I Introduction
Superresolution [78, 75, 76, 62, 72, 19, 73, 49, 74]
has a widerange of applications in various areas of imaging and computer vision, such as selflocalization
[54, 58, 26], image annotation [92, 93, 94, 90, 89, 91], surveillance [52, 53, 57], action recognition [87, 8, 6, 83, 82, 86, 9, 81, 7, 21, 88, 4], target tracking [85, 79, 64, 65], shape description and object recognition [106, 23, 22, 1, 1], imagebased rendering [2, 14, 12, 84], and camera motion estimation
[55, 26, 25, 24, 56, 28, 27, 5], to name a few. The goal of multiview superresolution image reconstruction is to obtain a high resolution (HR) image by fusing a sequence of degraded or aliased low resolution (LR) images of the same scene where degradation can be a consequence of motion, camera optics, atmosphere, insufficient sampling, etc.Approaches to solve the SRIR problem can be classified into frequency domain, interpolation, regularization, and learningbased methods (
[67, 96]).Fourierbased methods make use of the aliasing property of LR images in order to reconstruct an HR image. Even though these methods are intuitive and have low computational complexity, due to their global nature, they only allow linear space invariant blur (PSF). Moreover, it is difficult to identify a global frequencydomain a priori knowledge to overcome illposedness. Some examples of Fourier domain techniques include the following works: [98]
exploit the relationship between Continuous Fourier Transform (CFT) of the unknown HR scene and Discrete Fourier Transform (DFT) of the shifted and sampled LR images;
[69] utilize double random phase encoding in the imaging process in order to achieve SRIR with a single image. [70] apply combined Fourierwavelet deconvolution and denoising algorithm; and [100] perform registration and reconstruction in the Fourier domain using multiple unregistered images.Spatialdomain interpolationbased methods ([48, 109, 110]), on the other hand, tackle Fourierdomain obstacles by fusing the information from all LR images using a general interpolation technique, such as the nearest neighbor, bilinear, bicubic, etc. However, these methods result in overly smoothed images. As an example of interpolation based methods, [48] update the HR estimate by iteratively backprojecting the difference between the approximation and exact image. Moreover, [109] utilize multisurface fitting; whereas, [33] interpolate along the edge direction. [shekarforoush1999data] generalize Papoulis’ sampling theorem to merge nonuniform samples of multiple channels.
In order to stabilize the illposed problem of SRIR, regularizationbased methods optimize a cost function with a regularization term by incorporating prior knowledge. Some of these methods employ probabilistic estimators such as the maximum likelihood ([60, 37]), maximum a posteriori ([80, 18]), and Bayesian ([10, 20, 68]). The problems with these methods are determination of the prior model, high computational cost, and oversmoothing caused by regularization. More examples of regularizationbased methods include the following works: [63] employ Bregman iteration for Total Variation Regularization; [47] carry out canonical correlation analysis for human face SRIR task; [40] minimize norm and utilize regularization based on a bilateral prior; [39] minimize a multiterm cost function; and [105] constrain the SRIR process by using a spatially weighted TV model for different image regions. [62] investigate the contradiction between multichannel super resolution and regularization within the adaptive regularization framework.
Finally, machine learning is also used for SRIR (
[36, 43, 104]), where an HR image is obtained from LR images by utilizing training sets of LR/HR images or patch pairs. The problems with these methods again include their high computational cost, and correspondence ambiguities between HR and LR images. [103] utilize kernel partial least squares to implement regression along with a compensation of the primitive HR image with a residual HR image. Clustering and supervised neighbor embedding is employed by [107]. [45] combine multiimage SRIR and examplebased approaches based on the assumption that patches in natural images recur many times inside the image; [59]propose a video super resolution method with a convolutional neural network (CNN) that is trained on both spatial and temporal dimensions of videos; and
[31] extend video bilevel dictionary learning to multiframe super resolution using motion estimation.In order to overcome the drawbacks of aforementioned methods, recent research in SRIR explores waveletbased techniques ([50, 70, 34, 35, 97]). The intuition behind these approaches is that the LR images can be used to model the lowpass subbands of the unknown HR images, in order to reconstruct the high frequency information lost during image acquisition. The use of wavelets is also motivated by the fact that they are widely used in waveletencoded imaging ([3, 32, 71, 61]). Therefore, our motivation to study the proposed waveletbased SRIR is due to the drawbacks of aforementioned methods, growing trend of waveletencoded imaging, and the properties of wavelets which include orthogonality, signal localization, and low computing requirements.
Our approach is very much like singleimage superresolution methods that attempt to extend the spectrum of a single image to higher frequencies, but using information from other images, rather than prior knowledge. In that sense, our method is halfway between singleimage and multiimage methods, taking advantage of the best of both worlds. In the scope of this paper, we assume that displacements between the reference and other LR images are known a priori or estimated. We also assume that the displacements are pure translational, or corrected to be so. Our contributions are as follows: 1. We establish explicit closedform expressions that define how the local high frequency information that we aim to recover for the reference HR image is related to the local low frequency information in the given sequence of LR views. 2. We assume that the LR images correspond to the polyphase approximation coefficients of the first level wavelet transform of unknown HR images, which allows us to reduce the inverse problem to a collection of wellposed linear problems (local linearity). Our approach is closedform, and provides results that are superior to the stateoftheart. We provide the derived formulae utilizing the Haar wavelet transform as an example due to their locality and low computational requirements; however, a general formulation for wavelets can be derived as well. Our exceptional results are attributed to the accuracy, wellposedness, and the linearity of the equations derived in Section III, and the inherent nature of wavelets, making them very effective in signal localization.
The remainder of this paper is organized as follows. In Section II, a brief summary of waveletbased methods for SRIR is provided. The notations used throughout the paper along with the derived closedform linear relationships are defined in Section III. Section IV presents the general SRIR observation model, as well as the proposed approach. The stability analysis of the proposed SRIR method is examined in Section V. Finally, Section VI presents the experimental results and comparisons with both single and multiimage stateoftheart techniques.
Ii Related Work
Waveletbased SRIR approaches can be summarized as follows. In order to reduce noise in SRIR methods, [70] apply a combined Fourierwavelet deconvolution and denoising algorithm to multiframe SRIR. Authors first produce a sharp and noisy image by fast Fourier based image restoration, then reduce noise by space invariant nonlinear wavelet thresholding. The need to invert large matrices in their method results in solving the problem in the Fourier domain, which has its drawbacks mentioned in Section LABEL:intro. On the other hand, to reduce degradation artifacts such as blurring and the ringing effect, [95]
utilize zero padding in the wavelet domain followed by cycle spinning. Their method adopts a simplified edge profile and linear regression for edge degradations. Furthermore, to preserve edges,
[34] utilize stationary and discrete wavelet transforms together in an interpolationbased framework. However, even though better than conventional interpolation techniques, the latter two methods still lack sharp edges.[108] solve a constrained optimization problem utilizing wavelet domain Hidden Markov Tree (HMT) model for the prior knowledge problem, since HMT characterizes the statistics of real world images accurately. In order to suppress the artifacts left after employing their method, cycle spinning is used which leads to blurring as in other interpolation based methods.
[66], contrary to the conventional interpolation based methods, use the regularity and structure in the interlaced sampling of LR images. Even though, for 2D images, they utilize reshaping property of the Kronecker product, which only doubles the complexity for 1D, their method is based on conjugate gradient which is still time consuming. For deblurring, [29] derive iterative algorithms, which decompose HR image obtained from an iteration into different frequency components and add them to the next iteration. Their method utilizes wavelet thresholding for denoising, where highfrequency components are penalized, making their method dependent on accurate noise estimation. Moreover, [50] handle image registration and reconstruction together, by first estimating the homographies between multiple images, then reconstructing the HR image in a waveletbased iterative backprojection scheme.
Learningbased methods in waveletdomain include the following works: [51] handle the problem of representing the relationship between LRHR frames with training their dataset with HR images by learning from wavelet coefficients at finer scales, followed by regularization in a least squares manner; and [44]
follow Jiji’s method and employ discrete wavelet transform for training, where a cost function based on maximum a posteriori estimation is optimized with gradient descent method, employing an Inhomogeneous Gaussian Markov random field prior.
[35], to solve the problem of varying contents in different images or image patches, learn various sets of bases from a precollected dataset of example image patches, and select one set of bases adaptively to characterize the local sparse wavelet domain. However, these methods are all based on optimization which requires high computational cost.The above stated methods are performed either iteratively which requires high computational time or based on interpolation which results in overly smooth images. Our goal is to derive a direct relationship between LR images for a closedform SRIR solution, which prevents sacrificing high quality. Our paper can be viewed as a generalization of the work by [97] to some extent. [97] utilize Taylor series expansion to approximate the high frequency information that we want to recover. Their method is constrained to use LR images which have specific translations (namely 1 pixel in horizontal and 1 pixel in vertical directions). Our method, on the other hand, generalizes the translations for any shift; while determining the exact relationship between LR images and the subbands, instead of providing an approximation.
Iii Subpixel Shifts of a Low Resolution Image
Inband (i.e. wavelet domain) subpixel shift method along with the related notation are provided in this section.
Iiia Notation
Here, we provide the notations used throughout the paper in Table I.
Reference HR image  
level Haar wavelet transform approximation, horizontal, vertical, and diagonal detail coefficients of , respectively  
Matrices to be multiplied by approximation and detail coefficients (i.e. ) of the reference HR image, that are used to define inband shift of the reference LR image (i.e. )  
Number of hypothetically added levels in case of noninteger shifts  
Integer shift amount after the hypothetically added levels () 
Bold uppercase letters in the following sections demonstrate matrices whereas bold lowercase ones indicate vectors. The subscripts
demonstrate horizontal, vertical, and diagonal translations, respectively. Finally, the subscript indicates the th LR or HR image.IiiB Subpixel Shifts
Our goal for the proposed SRIR method is to reconstruct the lost high frequency information of an unknown HR image, given a sequence of subpixel shifted LR images. For this purpose, we first derive the relationship that relates these LR images to the high frequency information of the unknown HR image. This relationship can be described by inband shift (i.e. in the wavelet domain) of a reference LR image.
In order to find the aforementioned relationship, we first assume that the reference HR image is known. The reference LR image is the approximation coefficients obtained by decomposing the HR image for 1level Haar Transform. Then, we define shifted LR images based on the resultant Haar coefficients of the HR image. The shifting process is illustrated in Fig. 1, where shifted LR images (i.e. ) are described based on the first level approximation and detail coefficients of the reference HR image (i.e. ).
Below, we derive the mathematical expressions which demonstrate this relationship. The derived equations relate the highfrequency part (i.e. detail wavelet coefficients) of a reference HR image to the lowfrequency information provided by the LR image sequence.
Let A, H, V, and D be the first level approximation (i.e. reference LR image), horizontal, vertical, and diagonal detail coefficients, respectively, of a reference HR image, , of size , where and are positive integers. Since 1level wavelet transform reduces the size of HR image by half in each direction for approximation and detail coefficients, we require the size of HR image to be divisible by 2. Now, a translated LR image in an arbitrary direction can be expressed in matrix form using the level Haar transform of as in the following equation in (1).
(1) 
As already mentioned in Section IIIA, F and K stand for matrices to be multiplied by the first level lowpass and highpass subbands of the reference HR image, where subscripts and indicate horizontal and vertical shifts. stands for a shifted image in any direction. The low/highpass subbands together with are of size , and are , whereas and are .
By examining the translational shifts between two LR images in the Haar domain, we realize that horizontal translation reduces to zero and
to the identity matrix. This could be comprehended by examining the coefficient matrices defined later in this section (namely, Eq. (
3)), by making related vertical components zero (specifically, and ). This observation lets us define a horizontally shifted image by using only approximation and horizontal detail coefficients. Likewise, vertical translation solely necessitates approximation and vertical detail coefficients, in which case is reduced to zero and is equal to the identity matrix. As a result, the equation shown above in Eq. (1), can be expressed for each translation direction as in Eq. (2):(2)  
Here, our focus is on subpixel translations. Contrary to the general concept of approximating a subpixel shift by upsampling an image followed by an integer shift, our method models subpixel shift directly on the original coefficients of the reference HR image, without upsampling. We observe that:
(1) Viewing the wavelet transform of the image as in Fig. 2, upsampling an image is equivalent to adding levels to the bottom of the transform, and setting the detail coefficients to zero while the approximation coefficients remain the same.
(2) Shifting the upsampled image by an amount of is a counterpart of shifting the original image by an amount of , where is the number of added levels.
Fig. 2 demonstrates an example of the upsampling process described above where , which implies that only 1 level of zero detail coefficients are added. Assuming that the HR image is given, Haar Transform of this HR image can be found readily. For upsampling, these Haar Transform coefficients are utilized as approximation coefficients with more levels of detail coefficients which are set to be zero. Here, gray boxes demonstrate added zeros.
These observations allow us to shift a reference LR image inband (when the corresponding HR image is given) for a subpixel amount without actually upsampling it, which saves memory and reduces the computational cost. In order to shift the reference LR image, the original approximation and detail coefficients of the reference HR image are utilized with a hypothetically added level () and an integer shift value () at the added level.
Now, the aforementioned matrices, , , , and can be defined, in bidiagonal Toeplitz matrix form as follows.
(3) 
where and demonstrate the integer shift amounts at the hypothetically added level and the number of added levels for and directions, respectively.
As mentioned earlier, and are , while and are . Sizes of these matrices also indicate that inband shift of a reference LR image is performed using only the original level Haar coefficients (which are of size ) without upsampling. These matrices show that a 2pixel neighborhood in the approximation and detail coefficients of a reference HR image is utilized to shift a reference LR image inband. When the shift amount is negative, diagonals of the matrices interchange. We leave these matrices as square for them to be nonsingular in the SRIR process, otherwise these matrices could be adapted for periodic boundary condition by making them rectangular.
When the shift amount is not divisible by , in order to reach an integer value at the th level, the shift value at the original level is rounded to the closest decimal point which is divisible by .
Here, derived matrices to calculate the first level lowpass subband of a shifted image (i.e. ) are demonstrated. The counterparts for the first level detail coefficients (e.g. ) of the shifted HR images can be found in a similar manner, in order to shift the entire HR reference image directly in the Haar domain. Since for our SRIR method, we will only employ the relationship for approximation coefficients (i.e. LR images), we only provide the related equations.
Iv Super Resolution Image Reconstruction
In this section, we first present the SRIR observation model, followed by our proposed method.
Iva Observation Model
Let denote the desired HR image, and be the th observed LR image. The super resolution observation model is given by:
(4) 
where , , , and denote motion, blurring effect, downsampling operator, and noise term for the th LR image, respectively, and is the number of observed LR images. and i are the th LR image and unknown HR images, respectively, represented in lexicographical order.
Given a sequence of observed LR images, , the goal of SRIR is to reconstruct an unknown HR image, i.
IvB Proposed Method
As in the underlying idea of waveletbased SRIR algorithms, we assume that the given LR image sequence is the lowpass subbands (i.e. approximation coefficients of 1level Wavelet Transform) of unknown HR images. The goal is to reconstruct the unknown highpass subbands (i.e. detail coefficients of 1level Wavelet Transform) of one of these HR images which is chosen as the reference one. The SRIR method described below is the inverse process of the method described in Section III, where HR images are unknown, and high frequency information for one of these underlying HR images is estimated by solving a related linear system.
The relationship between two subpixel shifted LR images depends on the highpass subbands of the underlying reference HR image, as demonstrated in the previous section. This fact is used to construct a linear system of equations based on known LR images (i.e. in Section III) and unknown highpass subbands of the reference HR image (i.e. in Section III) using related formulae from Eq. (2) depending on the translation direction. Since there are three unknowns (i.e. horizontal, vertical, and diagonal detail coefficients of the unknown HR image), three shifted LR images together with the reference LR image are required to solve the linear system. Once this system is solved for the unknowns, inverse Haar transform utilizing the reference LR image and the estimated highpass subbands of the underlying unknown reference HR image gives the reconstructed HR image.
Fig. 3 shows a pictorial explanation of the proposed method, where solid boxes indicate known or estimated coefficients and dotted boxes show unknown ones. Images with the hat symbol (i.e. ) stands for estimated coefficients. As the figure demonstrates, assuming the LR sequence is first level approximation coefficients of the wavelet transform, we estimate the unknown high frequency information of the reference HR image in order to reconstruct the estimated HR image.
In the scope of this paper, we assume that the registration between images are known a priori or has been estimated. Translational shifts can be estimated using one of the methods by [42, 77, 41, 15, 16, 13, 17, 38, 101]. Even though the equations derived in Section III are for subpixel shifts, we apply the proposed SRIR method to the intersection area of any given shift, which may include an integer part, as well.
The proposed algorithm can also be explained step by step in Algorithm  Super Resolution Image Reconstruction as follows.
V Stability Analysis
In this section, we will investigate the stability of our method.
As mentioned in Section IV, our method constructs a linear system of equations based on given LR images and related shifts. Since the LR images (i.e. A, , , and ) and the displacements between them are known, , and D are the only unknowns of the constructed system. This linear system may appear in four forms which include:

1 horizontally, 1 vertically, 1 diagonally shifted image

1 horizontally, 2 diagonally shifted images

1 vertically, 2 diagonally shifted images

3 diagonally shifted images
along with the reference LR image, A, where second and third cases demonstrate the same properties. Thus, we will consider the first, second, and last cases in our analysis.
Case 1 (1 horizontal, 1 vertical, 1 diagonal) : This case constructs a linear system of equations exactly as shown in Eq. (2). This linear system is solved first for H using the equation for , then for V using the equation for , and finally for D using the equation for and substituting the information found for H and V. Since the coefficient matrices are invertible, this system is stable.
Case 2 and 3 (1 horizontal/vertical, 2 diagonal) : Here, we will explore Case 2 with 1 horizontally and 2 diagonally shifted images. Case 3 will demonstrate similar features as mentioned above.
This case includes one and two from Eq. (2) for one horizontal and two diagonally shifted images, where the linear system takes the form:
Again, as in Case 1, the first equation is stable, therefore H can be found easily. Solving equations for and for V results in:
(6) 
where
In order to tackle the instability problem caused by inverting multiplication and summation of matrices in Eq. (6), we right multiply this equation with . Since and differ only by the shift value for the two diagonally shifted images, results in the identity matrix multiplied by a scalar which depends only on the shifts. Thus, the equation for V becomes:
(7) 
where
where is defined as a constant.
Truncated Singular Value Decomposition (TSVD) is used with the resulting equation in (
7) to find V. Rank of TSVD method is decided based on minimizing the following cost function:(8) 
where U shows the identity (i.e. unit) matrix, subscript is Frobenius norm, and stands for rankr approximation of a matrix X.
In order to successfully truncate X at , we follow a theorem by [46] (Theorem 3.2), which implies that there must be a welldetermined gap between the two consecutive singular values at (i.e. ) and (i.e. ).
As one can see in Eq. (7), the stability of our method is partially dependent on the closeness of shift amounts.
Case 4 (3 diagonal) : The final case includes three diagonally shifted images together with the reference image. Therefore, the linear system is constructed as:
By solving the system above in Eq. (V) for H, we find a generalized Sylvester equation as in:
(10) 
where
R  
By examining for , in the generalized Sylvester equation, could be changed by multiplication by a scalar (as in in Eq. (7)), which leaves as an upper bidiagonal matrix, since is also upper bidiagonal. Moreover, since is an upper bidiagonal matrix, inverse of is an upper triangular matrix ([99]). Therefore, by multiplication of two upper triangular matrices, we obtain upper triangular matrices for . By following similar analysis, we observe that are lower triangular matrices.
Here, we refer to a theorem by [30] for a generalized Sylvester equation to have a unique solution. Interested reader can find the proof for this theorem in the referred paper; we include the theorem here to make this paper selfcontained.
Theorem: The matrix equation in (10) has a unique solution if and only if

and are regular matrix pencils,
and 
where
shows the generalized eigenvalues of the matrix pencils,
defines the spectra of the generalized eigenvalues, and demonstrates a matrix pencil.The matrix pencils constructed as , and , using the given and in Eq. (10), are not guaranteed either to be regular, or to have empty intersection of generalized eigenvalue spectra. For instance, when any of the two LR images have negative horizontal shift amount, the related has zero diagonals, and a zero element on the diagonal makes the matrix pencil singular when a matrix pencil is upper/lower triangular ([11]). Since we know that , and , for , are upper and lower triangular matrices, respectively, forming upper/lower triangular matrix pencils, two images with negative horizontal shifts satisfy requirements for singular matrix pencils.
Based on these facts, solution methods utilized for generalized Sylvester equation cannot be employed here. Therefore, in order to find a solution to the system in Eq. (V
), we first vectorize the equations using Kronecker tensor product, before solving for the unknowns:
for . Here, lowercase bold letters indicate columnvise vectorized versions of , and D, and these vectors have size . The Kronecker tensor products in parenthesis result in matrices of size , where and are the size of LR images.
By solving Eq. (V) for h, we find the following equation which appears similar to the equation for V in Eq. (6):
(12) 
where
for , and
FF  
FK  
KF  
KK 
Here, in order to solve for h, we follow a similar approach to the one used to reach Eq. (7) from Eq. (6), where Eq. (12) is left multiplied by in order to reduce the instability. Again, TSVD is utilized to solve the equation with the same cost function used in Eq. (8).
As in Cases 2 and 3, the stability of our solution depends partly on the closeness of shift values which affects the matrix inversions.
Vi Experimental Results
In this section, we first present the implementation details, followed by results for the proposed method along with comparisons to the recent stateoftheart and conventional techniques. Comparisons are made based on qualitative and quantitative evaluations on both commonly adapted test examples and real world images to demonstrate the influence of compression artifacts and sensor noise on the proposed method. LR image sequences are synthetically generated. Computational time efficiency of the proposed method against other methods are also presented. Moreover, HR and LR reference images for all test cases and zoomed parts in detailed areas for each image are provided.
Via Implementation Details
LR image sequences are synthesized by the method explained in Section IV (Imaging Process in Algorithm). LR images are divided into overlapping blocks of size , in order to reduce memory usage and decrease computational time.
To simulate the motion estimation error for the proposed method, HR reference image is shifted randomly for a shift amount which is not necessarily divisible by and shifts are rounded to the closest decimal divisible by for the calculations, as described before in Section III.
For the cases when the shift amounts are not subpixel (which might be integer or include an integer part), we find the intersection area of the images which can be described as subpixel shift. We apply the same method to the intersected area, where boundaries are lost for the maximum integer amount among all shifts.
In order to reduce the boundary problem caused by square coefficient matrices and which does not include the information in the boundaries, the last rows and columns of calculated , and D are extrapolated.
Color images are handled by the conventional approach ([35, 104]), in YIQ color space, where only the illuminance channel of images are dealt with the proposed method, since human visual system is more sensitive to changes in illuminance channel. The chrominance channels are upsampled using bicubic interpolation.
ViB Qualitative Comparison
We compared our method with both multiframe and single image SRIR techniques including interpolationbased ones which are Bicubic interpolation and Robust Super Resolution ([110]), a regularizationbased methods by [10] and [40], and finally a waveletdomain learningbased method by [35]. Compared methods were given the same input images and knowledge of registration (if required).
Figs. 4, 5, and 7 to 10 show results obtained with our method, the stateoftheart, and conventional ones. As can be seen from these figures, in zoomed areas particularly, the proposed method generates sharper edges with less artifacts compared to other methods.
While bicubic interpolation ((c) parts in all related figures) tend to introduce blur to the images, Robust SR technique leaves jaggy artifacts on the edges which are easily seen in (d) parts of Figs. 4, 5, and 7 of Lena, Car tag, and Resolution chart images. Babacan’s method alleviates the jaggy artifacts in most cases, yet the final results remain overly smoothed; whereas, Farsiu’s method does not reconstrut details that can be observed especially in part (f) of Figs. 7 and 8 of Resolution chart and Mandrill images. Dong’s method, even though better at removing artifacts and achieving natural looking results compared to the other methods, also is prone to leave blurry images as can be recognized without difficulty in (g) parts of Resolution chart, Mandrill and Car images in Figs. 7, 8, and 10. In addition, compared methods fail to recover fine details which is mostly recognized in car tags and numbers in Figs. 5, 7, 9, and 10.
On the other hand, the proposed method, seen in (h) parts of Figs. 4, 5, and 7 to 10, is able to recover sharp edges without visual artifacts, or blurring the images which leads to generating the closest results to the ground truth. Particularly, Resolution chart, Mandrill and Car images in Figs. 7, 8, and 10 demonstrate the high quality achieved with the proposed method. Recovered texture details with our method can be observed in all test cases upon a closer look, specifically in the feather texture of Lena’s hat in Fig. 4 and hair texture in cheeks of Mandrill image in Fig. 8. Overall, the proposed method removes artifacts and blur while preserving sharp edges without sacrificing a natural look.
ViC Quantitative Comparison
To further investigate the effectiveness of our method, we also conduct a comparison based on objective measurements PSNR, RMSE and SSIM ([102]), summarized in Table II. Comparisons are based on the illuminance channel of images reconstructed with all methods. The best of all cases are in bold text. While Robust SR performs the worst based on most measurements for all images except Circles in Fig. 9, Dong’s method has much better results compared to the other methods since in order to alleviate the correspondence ambiguity, their method uses different patches in a single image to learn various sets of bases. However, as can be seen also from Table II, in most cases, quantitative comparisons confirm visual ones which shows that our method outperforms the stateoftheart. Even though for Circles image in Fig. 9, the quantitative measurements are better for Dong’s method, with a closer look in the green bordered square, it can be seen that details are recovered better in our method, e.g. ”600” circle.
Image  Bicubic  [110]  [10]  [40]  [35]  Proposed  

PSNR  RMSE  SSIM  PSNR  RMSE  SSIM  PSNR  RMSE  SSIM  PSNR  RMSE  SSIM  PSNR  RMSE  SSIM  PSNR  RMSE  SSIM  
Lena  26.57  11.97  0.81  22.69  18.70  0.68  22.51  19.1  0.77  26.82  11.62  0.83  28.37  9.73  0.88  32.65  5.94  0.96 
Car tag  26.86  11.58  0.87  18.57  30.06  0.54  22.92  18.21  0.78  23.83  17.38  0.84  29.86  8.19  0.93  33.15  5.61  0.97 
Chart  24.28  15.57  0.84  19.80  26.10  0.72  22.00  20.26  0.79  23.00  18.05  0.85  26.07  12.68  0.89  29.85  8.20  0.96 
Mandrill  22.13  19.94  0.61  20.76  23.37  0.50  15.15  44.59  0.42  22.98  18.09  0.63  22.25  19.69  0.63  24.09  15.93  0.90 
Circles  14.15  50.02  0.66  22.55  19.02  0.96  25.23  13.97  0.97  27.42  10.85  0.98  37.33  3.47  0.99  33.57  5.34  0.99 
Car  24.01  16.07  0.80  21.43  21.64  0.71  21.47  21.54  0.74  23.23  17.59  0.82  25.02  14.30  0.86  31.14  7.07  0.97 
ViD Computational Efficiency
The computational complexity of the proposed method depends on matrix multiplications () along with the TSVD method ( where is the approximation rank). Since all blocks have the same size and use the same shift information, matrix inversions are handled only once, and the proposed super resolution method is applied to all blocks in parallel. Our method can also be applied as a sparse method in order to reduce time complexity, considering the fact that coefficient matrices are either bidiagonal or at most triangular matrices.
Time complexity of the proposed method and stateoftheart is compared in Table III, where average time taken for different size LR images is shown in seconds. Block size of all compared cases are set to . As can be seen from the table, the proposed method outperforms regularization based method ([10]
) (where outliers of
[10] are removed for a fair comparison) and learning based method ([35]) especially when the image sizes are relatively large. Since [104] learns a compact representation for image patch pairs in LR and HR images to capture the cooccurrence prior, their method has a lower computational time complexity for smaller size images; however, as the size increases, the proposed method outperforms [104] as well.A comparison of block sizes with time in seconds and PSNR for the proposed method is shown in Fig. 6. The results are calculated for 100 different images for 100 random shift amounts, and the average time and PSNR are shown in the graphs (after removing the outliers). As one can see from the graph, as block sizes increase, PSNR improves; however, time complexity increases at the same time. Therefore, the block sizes can be decided based on the application depending on the importance of time or accuracy. Although the graph demonstrates the results for square sized blocks, the block sizes are decided based on the image sizes, which can as well be rectangular.
Vii Summary and Conclusions
As a final remark, a direct waveletbased super resolution technique is proposed in this paper by first deriving exact inband relationships between two subpixel shifted images, then utilizing these relationships in a linear system form to reconstruct high frequency information of a low resolution reference image. Our results outperform the conventional as well as advanced recently published methods. We attribute this to the accuracy, wellposedness and the linearity of the equations derived in Section III and the inherent local nature of wavelets, making them very effective in signal localization. In summary, we present herein a method for superresolution by effectively estimating the high frequency information in the Haar domain, which in a sense is a hybrid approach between single image and multiimage methods, taking advantage of the best of both worlds.
References
 [1] Muhammad Ali and Hassan Foroosh. Character recognition in natural scene images using rank1 tensor decomposition. In Image Processing (ICIP), 2016 IEEE International Conference on, pages 2891–2895. IEEE, 2016.
 [2] Mais Alnasser and Hassan Foroosh. Imagebased rendering of synthetic diffuse objects in natural scenes. In Pattern Recognition, 2006. ICPR 2006. 18th International Conference on, volume 4, pages 787–790. IEEE.
 [3] Marc Antonini, Michel Barlaud, Pierre Mathieu, and Ingrid Daubechies. Image coding using wavelet transform. IEEE Transactions on image processing, 1(2):205–220, 1992.
 [4] Nazim Ashraf and Hassan Foroosh. Human action recognition in video data using invariant characteristic vectors. In Image Processing (ICIP), 2012 19th IEEE International Conference on, pages 1385–1388. IEEE, 2012.
 [5] Nazim Ashraf, Imran Junejo, and Hassan Foroosh. Nearoptimal mosaic selection for rotating and zooming video cameras. Computer Vision–ACCV 2007, pages 63–72, 2007.
 [6] Nazim Ashraf, Yuping Shen, Xiaochun Cao, and Hassan Foroosh. Viewinvariant action recognition using weighted fundamental ratios. Journal of Computer Vision and Image Understanding (CVIU), 117:587–602, 2013.
 [7] Nazim Ashraf, Yuping Shen, and Hassan Foroosh. Viewinvariant action recognition using rank constraint. In Pattern Recognition (ICPR), 2010 20th International Conference on, pages 3611–3614. IEEE, 2010.
 [8] Nazim Ashraf, Chuan Sun, and Hassan Foroosh. Viewinvariant action recognition using projective depth. Journal of Computer Vision and Image Understanding (CVIU), 123:41–52, 2014.
 [9] Nazim Ashraf, Chuan Sun, and Hassan Foroosh. View invariant action recognition using projective depth. Computer Vision and Image Understanding, 123:41–52, 2014.
 [10] S Derin Babacan, Rafael Molina, and Aggelos K Katsaggelos. Variational bayesian super resolution. Image Processing, IEEE Transactions on, 20(4):984–999, 2011.
 [11] Zhaojun Bai, James Demmel, Jack Dongarra, Axel Ruhe, and Henk van der Vorst. Templates for the solution of algebraic eigenvalue problems: a practical guide, volume 11. Siam, 2000.
 [12] Murat Balci, Mais Alnasser, and Hassan Foroosh. Imagebased simulation of gaseous material. In Image Processing, 2006 IEEE International Conference on, pages 489–492. IEEE, 2006.
 [13] Murat Balci, Mais Alnasser, and Hassan Foroosh. Subpixel alignment of mri data under cartesian and logpolar sampling. In Pattern Recognition, 2006. ICPR 2006. 18th International Conference on, volume 3, pages 607–610. IEEE, 2006.
 [14] Murat Balci and Hassan Foroosh. Realtime 3d fire simulation using a springmass model. In MultiMedia Modelling Conference Proceedings, 2006 12th International, pages 8–pp. IEEE.
 [15] Murat Balci and Hassan Foroosh. Estimating subpixel shifts directly from the phase difference. In Image Processing, 2005. ICIP 2005. IEEE International Conference on, volume 1, pages I–1057. IEEE, 2005.
 [16] Murat Balci and Hassan Foroosh. Inferring motion from the rank constraint of the phase matrix. In Acoustics, Speech, and Signal Processing, 2005. Proceedings.(ICASSP’05). IEEE International Conference on, volume 2, pages ii–925. IEEE, 2005.
 [17] Murat Balci and Hassan Foroosh. Subpixel registration directly from the phase difference. EURASIP Journal on Applied Signal Processing, 2006:231–231, 2006.
 [18] Stefanos P Belekos, Nikolas P Galatsanos, and Aggelos K Katsaggelos. Maximum a posteriori video superresolution using a new multichannel image prior. Image Processing, IEEE Transactions on, 19(6):1451–1464, 2010.
 [19] M Berthod, M Werman, H Shekarforoush, and J Zerubia. Refining depth and luminance information using superresolution. In Computer Vision and Pattern Recognition, pages 654–657, 1994.
 [20] Christopher M Bishop and Michael E Tipping. Bayesian image super resolution, September 12 2006. US Patent 7,106,914.
 [21] Hakan Boyraz, Syed Zain Masood, Baoyuan Liu, Marshall Tappen, and Hassan Foroosh. Action recognition by weaklysupervised discriminative region localization.
 [22] Ozan Cakmakci, Brendan Moore, Hassan Foroosh, and Jannick Rolland. Optimal local shape description for rotationally nonsymmetric optical surface design and analysis. Optics Express, 16(3):1583–1589, 2008.

[23]
Ozan Cakmakci, Sophie Vo, Hassan Foroosh, and Jannick Rolland.
Application of radial basis functions to shape description in a dualelement offaxis magnifier.
Optics Letters, 33(11):1237–1239, 2008.  [24] Xiaochun Cao and Hassan Foroosh. Camera calibration without metric information using 1d objects. In Image Processing, 2004. ICIP’04. 2004 International Conference on, volume 2, pages 1349–1352. IEEE, 2004.
 [25] Xiaochun Cao and Hassan Foroosh. Camera calibration using symmetric objects. IEEE Transactions on Image Processing, 15(11):3614–3619, 2006.
 [26] Xiaochun Cao and Hassan Foroosh. Camera calibration and light source orientation from solar shadows. Journal of Computer Vision & Image Understanding (CVIU), 105:60–72, 2007.
 [27] Xiaochun Cao, Jiangjian Xiao, and Hassan Foroosh. Camera motion quantification and alignment. In Pattern Recognition, 2006. ICPR 2006. 18th International Conference on, volume 2, pages 13–16. IEEE.
 [28] Xiaochun Cao, Jiangjian Xiao, and Hassan Foroosh. Selfcalibration using constant camera motion. In Pattern Recognition, 2006. ICPR 2006. 18th International Conference on, volume 1, pages 595–598. IEEE.
 [29] Raymond H Chan, Tony F Chan, Lixin Shen, and Zuowei Shen. Wavelet algorithms for highresolution image reconstruction. SIAM Journal on Scientific Computing, 24(4):1408–1432, 2003.
 [30] Kingwah Eric Chu. The solution of the matrix equations axb cxd= e and (ya dz, yc bz)=(e, f). Linear Algebra and its Applications, 93:93–105, 1987.
 [31] Qiqin Dai, Seunghwan Yoo, Armin Kappeler, and Aggelos K Katsaggelos. Sparse representationbased multiple frame video superresolution. IEEE Transactions on Image Processing, 26(2):765–781, 2017.
 [32] Geoffrey M Davis and Aria Nosratinia. Waveletbased image coding: an overview. In Applied and computational control, signals, and circuits, pages 369–434. Springer, 1999.
 [33] Osborn de Lima, Sreenath R Vantaram, Sankaranarayanan Piramanayagam, Eli Saber, and Kurt R Bengtson. An edge directed super resolution technique for multimedia applications. In IS&T/SPIE Electronic Imaging, pages 86630F–86630F. International Society for Optics and Photonics, 2013.
 [34] Hasan Demirel and Gholamreza Anbarjafari. Image resolution enhancement by using discrete and stationary wavelet decomposition. IEEE TIP, 20(5):1458–1460, 2011.
 [35] Weisheng Dong, D Zhang, Guangming Shi, and Xiaolin Wu. Image deblurring and superresolution by adaptive sparse domain selection and adaptive regularization. Image Processing, IEEE Transactions on, 20(7):1838–1857, 2011.
 [36] Michael Elad and Dmitry Datsenko. Examplebased regularization deployed to superresolution reconstruction of a single image. The Computer Journal, 52(1):15–30, 2009.
 [37] Michael Elad and Yacov HelOr. A fast superresolution reconstruction algorithm for pure translational motion and common spaceinvariant blur. Image Processing, IEEE Transactions on, 10(8):1187–1193, 2001.
 [38] Georgios D Evangelidis and Emmanouil Z Psarakis. Parametric image alignment using enhanced correlation coefficient maximization. IEEE Transactions on Pattern Analysis and Machine Intelligence, 30(10):1858–1865, 2008.
 [39] Sina Farsiu, Michael Elad, and Peyman Milanfar. Multiframe demosaicing and superresolution of color images. Image Processing, IEEE Transactions on, 15(1):141–159, 2006.
 [40] Sina Farsiu, M Dirk Robinson, Michael Elad, and Peyman Milanfar. Fast and robust multiframe super resolution. IEEE transactions on image processing, 13(10):1327–1344, 2004.
 [41] Hassan Foroosh and Murat Balci. Subpixel registration and estimation of local shifts directly in the fourier domain. In Image Processing, 2004. ICIP’04. 2004 International Conference on, volume 3, pages 1915–1918. IEEE, 2004.
 [42] Hassan Foroosh, Josiane B Zerubia, and Marc Berthod. Extension of phase correlation to subpixel registration. IEEE transactions on image processing, 11(3):188–200, 2002.
 [43] William T Freeman, Thouis R Jones, and Egon C Pasztor. Examplebased superresolution. Computer Graphics and Applications, IEEE, 22(2):56–65, 2002.
 [44] Prakash P Gajjar and Manjunath V Joshi. New learning based superresolution: use of dwt and igmrf prior. Image Processing, IEEE Transactions on, 19(5):1201–1213, 2010.
 [45] Daniel Glasner, Shai Bagon, and Michal Irani. Superresolution from a single image. In Computer Vision, 2009 IEEE 12th International Conference on, pages 349–356. IEEE, 2009.
 [46] Per Christian Hansen. The truncated svd as a method for regularization. BIT Numerical Mathematics, 27(4):534–553, 1987.
 [47] Hua Huang, Huiting He, Xin Fan, and Junping Zhang. Superresolution of human face image using canonical correlation analysis. Pattern Recognition, 43(7):2532–2543, 2010.
 [48] Michal Irani and Shmuel Peleg. Improving resolution by image registration. CVGIP, 53(3):231–239, 1991.
 [49] Apurva Jain, Supraja Murali, Nicolene Papp, Kevin Thompson, Kyesung Lee, Panomsak Meemon, Hassan Foroosh, and Jannick P Rolland. Superresolution imaging combining the design of an optical coherence microscope objective with liquidlens based dynamic focusing capability and computational methods. In Optical Engineering+ Applications, pages 70610C–70610C. International Society for Optics and Photonics, 2008.
 [50] Hui Ji and Cornelia Fermuller. Robust waveletbased superresolution reconstruction: theory and algorithm. IEEE PAMI, 31(4):649–660, 2009.
 [51] CV Jiji, Manjunath V Joshi, and Subhasis Chaudhuri. Singleframe image superresolution using learned wavelet coefficients. International journal of Imaging systems and Technology, 14(3):105–112, 2004.
 [52] Imran Junejo, Xiaochun Cao, and Hassan Foroosh. Autoconfiguration of a dynamic nonoverlapping camera network. IEEE Trans. Systems, Man, and Cybernetics, 37(4):803–816, 2007.
 [53] Imran Junejo and Hassan Foroosh. Euclidean path modeling for video surveillance. Image and Vision Computing (IVC), 26(4):512–528, 2008.
 [54] Imran Junejo and Hassan Foroosh. Gps coordinates estimation and camera calibration from solar shadows. Computer Vision and Image Understanding (CVIU), 114(9):991–1003, 2010.
 [55] Imran Junejo and Hassan Foroosh. Optimizing ptz camera calibration from two images. Machine Vision and Applications (MVA), pages 1–15, 2011.
 [56] Imran N Junejo, Xiaochun Cao, and Hassan Foroosh. Calibrating freely moving cameras. In Pattern Recognition, 2006. ICPR 2006. 18th International Conference on, volume 4, pages 880–883. IEEE.
 [57] Imran N. Junejo and Hassan Foroosh. Trajectory rectification and path modeling for video surveillance. In Computer Vision, 2007. ICCV 2007. IEEE 11th International Conference on, pages 1–7. IEEE, 2007.
 [58] Imran N Junejo and Hassan Foroosh. Gps coordinate estimation from calibrated cameras. In Pattern Recognition, 2008. ICPR 2008. 19th International Conference on, pages 1–4. IEEE, 2008.
 [59] Armin Kappeler, Seunghwan Yoo, Qiqin Dai, and Aggelos K Katsaggelos. Video superresolution with convolutional neural networks. IEEE Transactions on Computational Imaging, 2(2):109–122, 2016.
 [60] Kyungsang Kim, Junhong Min, Lina Carlini, Michael Unser, Suliana Manley, Daejong Jeon, and J Ye. Fast maximum likelihood highdensity lowsnr superresolution localization microscopy. In 10th International Conference on Sampling Theory and Applications: Invited Session VII: Sampling in Bio Imaging, Jacobs University, Bremen, Gemerny, volume 3, 2013.
 [61] Zheng Liu, Brian Nutter, and Sunanda Mitra. Compressive sampling in fast waveletencoded mri. In SSIAI, 2012 IEEE Southwest Symposium on, pages 137–140.
 [62] Anne Lorette, Hassan Shekarforoush, and Josiane Zerubia. Superresolution with adaptive regularization. In Image Processing, 1997. Proceedings., International Conference on, volume 1, pages 169–172. IEEE, 1997.
 [63] Antonio Marquina and Stanley J Osher. Image superresolution by tvregularization and bregman iteration. Journal of Scientific Computing, 37(3):367–382, 2008.
 [64] Brian Milikan, Aritra Dutta, Qiyu Sun, and Hassan Foroosh. Compressed infrared target detection using stochastically trained least squares. IEEE Transactions on Aerospace and Electronics Systems, page accepted, 2017.
 [65] Brian Millikan, Aritra Dutta, Nazanin Rahnavard, Qiyu Sun, and Hassan Foroosh. Initialized iterative reweighted least squares for automatic target recognition. In Military Communications Conference, MILCOM 20152015 IEEE, pages 506–510. IEEE, 2015.
 [66] Nhat Nguyen and Peyman Milanfar. An efficient waveletbased algorithm for image superresolution. In ICIP, volume 2, pages 351–354. IEEE, 2000.
 [67] Sung Cheol Park, Min Kyu Park, and Moon Gi Kang. Superresolution image reconstruction: a technical overview. IEEE Signal Processing Magazine, 20(3):21–36, 2003.
 [68] Lyndsey C Pickup, David P Capel, Stephen J Roberts, and Andrew Zisserman. Overcoming registration uncertainty in image superresolution: maximize or marginalize? EURASIP Journal on Advances in Signal Processing, 2007(2):20–20, 2007.
 [69] Yair Rivenson, Adrian Stern, and Bahram Javidi. Single exposure superresolution compressive imaging by double phase encoding. Optics express, 18(14):15094–15103, 2010.
 [70] M Dirk Robinson, Cynthia A Toth, Joseph Y Lo, and Sina Farsiu. Efficient fourierwavelet superresolution. Image Processing, IEEE Transactions on, 19(10):2669–2681, 2010.
 [71] Hacene Serrai and Lotfi Senhadji. Acquisition time reduction in magnetic resonance spectroscopic imaging using discrete wavelet encoding. JMR, 177(1):22–30, 2005.
 [72] H Shekarforoush, R Chellappa, H Niemann, H Seidel, and B Girod. Multichannel superresolution for images sequences with applications to airborne video data. Proc. of IEEE Image and Multidimensional Digital Signal Processing, pages 207–210, 1998.
 [73] Hassan Shekarforoush. Conditioning bounds for multiframe superresolution algorithms. Computer Vision Laboratory, Center for Automation Research, University of Maryland, 1999.
 [74] Hassan Shekarforoush, Amit Banerjee, and Rama Chellappa. Super resolution for fopen sar data. In AeroSense’99, pages 123–129. International Society for Optics and Photonics, 1999.
 [75] Hassan Shekarforoush, Marc Berthod, Michael Werman, and Josiane Zerubia. Subpixel bayesian estimation of albedo and height. International Journal of Computer Vision, 19(3):289–300, 1996.
 [76] Hassan Shekarforoush, Marc Berthod, and Josiane Zerubia. 3d superresolution using generalized sampling expansion. In Image Processing, 1995. Proceedings., International Conference on, volume 2, pages 300–303. IEEE, 1995.
 [77] Hassan Shekarforoush, Marc Berthod, and Josiane Zerubia. Subpixel image registration by estimating the polyphase decomposition of cross power spectrum. In Computer Vision and Pattern Recognition, 1996. Proceedings CVPR’96, 1996 IEEE Computer Society Conference on, pages 532–537. IEEE, 1996.
 [78] Hassan Shekarforoush and Rama Chellappa. Datadriven multichannel superresolution with application to video sequences. Journal of Optical Society of AmericaA, 16(3):481–492, 1999.
 [79] Hassan Shekarforoush and Rama Chellappa. A multifractal formalism for stabilization, object detection and tracking in flir sequences. In Image Processing, 2000. Proceedings. 2000 International Conference on, volume 3, pages 78–81. IEEE, 2000.
 [80] Huanfeng Shen, Liangpei Zhang, Bo Huang, and Pingxiang Li. A map approach for joint motion estimation, segmentation, and super resolution. TIP, 16(2):479–490, 2007.
 [81] Yuping Shen, Nazim Ashraf, and Hassan Foroosh. Action recognition based on homography constraints. In Pattern Recognition, 2008. ICPR 2008. 19th International Conference on, pages 1–4. IEEE, 2008.
 [82] Yuping Shen and Hassan Foroosh. Viewinvariant action recognition using fundamental ratios. In Computer Vision and Pattern Recognition, 2008. CVPR 2008. IEEE Conference on, pages 1–6. IEEE, 2008.
 [83] Yuping Shen and Hassan Foroosh. Viewinvariant action recognition from point triplets. IEEE Transactions on Pattern Analysis and Machine Intelligence (PAMI), 31(10):1898–1905, 2009.
 [84] Yuping Shen, Fei Lu, Xiaochun Cao, and Hassan Foroosh. Video completion for perspective camera under constrained motion. In Pattern Recognition, 2006. ICPR 2006. 18th International Conference on, volume 3, pages 63–66. IEEE.
 [85] Chen Shu, Luming Liang, Wenzhang Liang, and Hassan Forooshh. 3d pose tracking with multitemplate warping and sift correspondences. IEEE Trans. on Circuits and Systems for Video Technology, 26(11):2043–2055, 2016.
 [86] Chuan Sun, Imran Junejo, and Hassan Foroosh. Action recognition using rank1 approximation of joint selfsimilarity volume. In Computer Vision (ICCV), 2011 IEEE International Conference on, pages 1007–1012. IEEE, 2011.
 [87] Chuan Sun, Imran Junejo, Marshall Tappen, and Hassan Foroosh. Exploring sparseness and selfsimilarity for action recognition. IEEE Transactions on Image Processing, 24(8):2488–2501, 2015.
 [88] Chuan Sun, Marshall Tappen, and Hassan Foroosh. Featureindependent action spotting without human localization, segmentation or framewise tracking. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 2689–2696, 2014.
 [89] Amara Tariq and Hassan Foroosh. Scenebased automatic image annotation. In Image Processing (ICIP), 2014 IEEE International Conference on, pages 3047–3051. IEEE, 2014.
 [90] Amara Tariq and Hassan Foroosh. Featureindependent context estimation for automatic image annotation. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 1958–1965, 2015.
 [91] Amara Tariq and Hassan Foroosh. Tclustering: Image clustering by tensor decomposition. In Image Processing (ICIP), 2015 IEEE International Conference on, pages 4803–4807. IEEE, 2015.
 [92] Amara Tariq, Asim Karim, and Hassan Foroosh. A contextdriven extractive framework for generating realistic image descriptions. IEEE Transactions on Image Processing, 26(2):619–632, 2002.
 [93] Amara Tariq, Asim Karim, and Hassan Foroosh. Nelasso: Building named entity relationship networks using sparse structured learning. IEEE Trans. on on Pattern Analysis and Machine Intelligence, page accepted, 2017.
 [94] Amara Tariq, Asim Karim, Fernando Gomez, and Hassan Foroosh. Exploiting topical perceptions over multilingual text for hashtag suggestion on twitter. In The TwentySixth International FLAIRS Conference, 2013.
 [95] A Temizel and T Vlachos. Wavelet domain image resolution enhancement using cyclespinning. Electronics Letters, 41(3):119–121, 2005.
 [96] Jing Tian and KaiKuang Ma. A survey on superresolution imaging. Signal, Image and Video Processing, 5(3):329–342, 2011.
 [97] CS Tong and KT Leung. Superresolution reconstruction based on linear interpolation of wavelet coefficients. Multidimensional Systems and Signal Processing, 18(23):153–171, 2007.
 [98] RY Tsai and Thomas S Huang. Multiframe image restoration and registration. Advances in computer vision and Image Processing, 1(2):317–339, 1984.
 [99] Raf Vandebril, Marc Van Barel, and Nicola Mastronardi. Matrix computations and semiseparable matrices: Linear systems (volume 1), 2007.
 [100] Patrick Vandewalle, Luciano Sbaiz, Joos Vandewalle, and Martin Vetterli. Superresolution from unregistered and totally aliased signals using subspace methods. Signal Processing, IEEE Transactions on, 55(7):3687–3703, 2007.
 [101] Patrick Vandewalle, Sabine Süsstrunk, and Martin Vetterli. A frequency domain approach to registration of aliased images with application to superresolution. EURASIP Journal on applied signal processing, 2006:233–233, 2006.
 [102] Zhou Wang, Alan C Bovik, Hamid R Sheikh, and Eero P Simoncelli. Image quality assessment: from error visibility to structural similarity. TIP, 13(4):600–612, 2004.
 [103] Wei Wu, Zheng Liu, and Xiaohai He. Learningbased super resolution using kernel partial least squares. Image and Vision Computing, 29(6):394–406, 2011.
 [104] Jianchao Yang, John Wright, Thomas S Huang, and Yi Ma. Image superresolution via sparse representation. Image Processing, IEEE Transactions on, 19(11):2861–2873, 2010.
 [105] Qiangqiang Yuan, Liangpei Zhang, and Huanfeng Shen. Multiframe superresolution employing a spatially weighted total variation model. IEEE Transactions on circuits and systems for video technology, 22(3):379–392, 2012.
 [106] Changqing Zhang, Xiaochun Cao, and Hassan Foroosh. Constrained multiview video face clustering. IEEE Transactions on Image Processing, 24(11):4381–4393, 2015.
 [107] Kaibing Zhang, Xinbo Gao, Xuelong Li, and Dacheng Tao. Partially supervised neighbor embedding for examplebased image superresolution. Selected Topics in Signal Processing, IEEE Journal of, 5(2):230–239, 2011.
 [108] Shubin Zhao, Hua Han, and Silong Peng. Waveletdomain hmtbased image superresolution. In ICIP, volume 2, pages II–953. IEEE, 2003.
 [109] Fei Zhou, Wenming Yang, and Qingmin Liao. Interpolationbased image superresolution using multisurface fitting. IEEE Transactions on Image Processing, 21(7):3312–3318, 2012.
 [110] Assaf Zomet, Alex RavAcha, and Shmuel Peleg. Robust superresolution. In CVPR, volume 1, pages I–645. IEEE, 2001.
Comments
There are no comments yet.