Computational approaches to the analysis of collective behavior in social insects increasingly rely on motion paths as an intermediate data layer from which one can infer individual behaviors or social interactions. Honey bees are a popular model for learning and memory. Previous experience has been shown to affect and modulate future social interactions. So far, no lifetime history observations have been reported for all bees of a colony. In a previous work we introduced a recording setup customized to track up to 4000 marked bees over several weeks. Due to detection and decoding errors of the bee markers, linking the correct correspondences through time is non-trivial. In this contribution we present an in-depth description of the underlying multi-step algorithm which produces motion paths, and also improves the marker decoding accuracy significantly. The proposed solution employs two classifiers to predict the correspondence of two consecutive detections in the first step, and two tracklets in the second. We automatically tracked ~2000 marked honey bees over 10 weeks with inexpensive recording hardware using markers without any error correction bits. We found that the proposed two-step tracking reduced incorrect ID decodings from initially ~13% to around 2% post-tracking. Alongside this paper, we publish the first trajectory dataset for all bees in a colony, extracted from ~3 million images covering three days. We invite researchers to join the collective scientific effort to investigate this intriguing animal system. All components of our system are open-source.
Social insect colonies are popular model organisms for self-organization and collective decision making. Devoid of central control, it often appears miraculous how orderly termites build their nests or ant colonies organize their labor. Honey bees are a particularly popular example - they stand out due to a rich repertoire of communication behaviors (von Frisch, 1965; Seeley, 2010) and their highly flexible division of labor (Robinson, 1992; Brian R. Johnson, 2010). A honey bee colony robustly adapts to changing conditions, whether it may be a hole in the hive that needs to be repaired, intruders that need to be fended off, brood that needs to be reared, or food that needs to be found and processed. The colony behavior emerges from interactions of many thousand individuals. The complexity that results from the vast number of individuals is increased by the fact that bees are excellent learners: empirical evidence indicates that personal experience can modulate communication behavior (Grüter et al., 2006; Balbuena et al., 2012; Grüter and Farina, 2009; Grüter and Ratnieks, 2011; Goyret and Farina, 2005; De Marco and Farina, 2001; Richter and Waddington, 1993). Especially among foragers, personal experience may be very variable. The various locations a forager visits might be dispersed over large distances (up to several kilometers around the hive) and each site might offer different qualities of food, or even pose threats. Thus, no two individuals share the same history and experiences. Evaluating how personal experience shapes the emergence of collective behavior and how individual information is communicated to and processed by the colony requires robust identification of individual bees over long time periods.
However, insects are particularly hard to distinguish by a human observer. Tracking a bee manually is therefore difficult to realize without marking these animals individually. Furthermore, following more than one individual simultaneously is almost impossible for the human eye. Thus, the video recording must be watched once per individual, which, in the case of a bee hive, might be several hundred or thousand times. Processing long time spans or the observation of many bees is therefore highly infeasible, or is limited to only a small group of animals. Most studies furthermore focused on one focal property, such as certain behaviors or the position of the animal. Over the last decades, various aspects of the social interactions in honey bee colonies have been investigated with remarkable efforts in data collection: Naug (Naug, 2008) manually followed around 1000 marked bees in a one hour long video to analyze food exchange interactions. Barachi and Cini (Baracchi and Cini, 2014) manually extracted the positions of 211 bees once per minute for 10 hours of video data to analyze the colony's proximity network. Biesmeijer and Seeley (Biesmeijer and Seeley, 2005) observed foraging related behaviours of a total of 120 marked bees over 20 days. Couvillon and coworkers manually decoded over 5000 waggle dances from video (Couvillon et al., 2014). Research questions requiring multiple properties, many individuals, or long time frames are limited by the costs of manual labor.
In recent years, computer vision software for the automatic identification and tracking of animals has evolved into a popular tool for quantifying behavior(Krause et al., 2013; Dell et al., 2014a). Although some focal behaviors might be extracted from the video feed directly (Berman et al., 2014; Wiltschko et al., 2015; Wario et al., 2017), tracking the position of an animal often suffices to infer its behavioral state (Kabra et al., 2013; Eyjolfsdottir et al., 2016; Blut et al., 2017). Tracking bees within a colony is a particularly challenging task due to dense populations, similar target appearance, frequent occlusions and a significant portion of the colony frequently leaving the hive. The exploration flights of foragers might take several hours, guard bees might stay outside the entire day to inspect incoming individuals. The observation of individual activity over many weeks, hence, requires robust means for unique identification.
For a system that robustly decodes the identity of a given detection, the tracking task reduces to simply connecting matching IDs. Recently, three marker-based insect tracking systems (Mersch et al., 2013; Crall et al., 2015; Gernat et al., 2018) have been proposed that use a binary code with up to 26 bits for error correction (Thompson, 1983). The decoding process can reliably detect and correct errors, or, reject a detection that can not be decoded. There are two disadvantages to this approach. First, error correction requires relatively expensive recording equipment (most systems use at least a 20 MP sensor with a high quality lens). Second, detections that could not be decoded can usually not be integrated into the trajectory, effectively reducing the detection accuracy and sample rate.
In contrast to these solutions, we have developed a system called BeesBook that uses much less expensive recording equipment (Wario et al., 2015). Figure 1 shows our recording setup, Figure 2 visualizes the processing steps performed after the recording. Our system localizes tags with a recall of 98% at 99% precision and decodes 86% IDs correctly without relying on error correcting codes (Wild et al., 2018). See Figure 3 for the tag design. Linking detections only based on matching IDs would quickly accumulate errors, long-term trajectories would exhibit gaps or jumps between individuals. Following individuals robustly, thus, requires a more elaborate tracking algorithm.
The field of multiple object tracking has produced numerous solutions to various use-cases such as pedestrian and vehicle tracking (for reviews see Cox (1993); Wu et al. (2013); Luo et al. (2014); Betke and Wu (2016)). Animals, especially insects, are harder to distinguish and solutions for tracking multiple animals over long time frames are far less numerous (see Dell et al. (2014b) for a review on animal tracking). Since our target subjects may leave the area under observation at any time, the animal’s identity cannot be preserved by tracking alone. We require some means of identification for a new detection, whether it be paint marks or number tags on the animals, or identity-preserving descriptors extracted from the detection.
While color codes are infeasible with monochromatic imaging, using image statistics to “fingerprint” sequences of visible animals (Pérez-Escudero et al., 2014; Kühl and Burghardt, 2013; Wang and Yeung, 2013) may work even with unstructured paint markers. Merging tracklets after occlusions can then be done by matching fingerprints. However, it remains untested whether these approaches can resolve the numerous ambiguities in long-term observations of many hundreds or thousands of bees that may leave the hive for several hours.
In the following, we describe the features that we used to train machine learning classifiers to link individual detections and short tracklets in a crowded bee hive. We evaluate our results with respect to path and ID correctness. We conclude that long-term tracking can be performed without marker-based error correction codes. Tracking can, thus, be conducted without expensive high-resolution, low-noise camera equipment. Instead, decoding errors in simple markers can be mitigated by the proposed tracking solution, leading to a higher final accuracy of the assigned IDs compared to other marker-based systems that do not employ a tracking step.
2 Description of Methods
2.1 Problem statement and overview of tracking approach
The tracking problem is defined as follows: Given a set of detections (timestamp, location, orientation and ID information), find correct correspondences among detections over time (tracks) and assign the correct ID to each track. The ID information of the detections can contain errors. Additionally, correct correspondences between detections of consecutive frames might not exist due to missing detections caused by occluded markers. In our dataset, the ID information consists of a number in the range of 0 to 4095, represented by 12 bits. Each bit is given as a value between 0.0 and 1.0 which corresponds to the probability that the bit is set.
To solve the described tracking problem, we propose an iterative tracking approach, similar to previous works (for reviews, see Betke and Wu (2016); Luo et al. (2014)). We use two steps: 1. Consecutive detections are combined into short but reliable tracklets (Rosemann., 2017). 2. These tracklets are connected over longer gaps (Boenisch, 2017). Previous work employing machine learning mostly scored different distance measures separately to combine them into one thresholded value for the first tracking step (Wu and Nevatia, 2007; Huang et al., 2008; Wang et al., 2014; Fasciano et al., 2013). For merging longer tracks, boosting models to predict a ranking between candidate tracklets have been proposed (Huang et al., 2008; Fasciano et al., 2013). We use machine learning models in both steps to learn the probability that two detections, or tracklets, correspond. We train the models on a manually labeled dataset of ground truth tracklets. The features that are used to predict correspondence can differ between detection level and tracklet level, so we treat these two stages as separate learning problems. Both of our tracking steps use the Hungarian algorithm (Kuhn, 1955) to assign likely matches between detections in subsequent time steps based on the predicted probability of correspondence. In the following, we describe which features are suitable for each step and how we used various regression models to create accurate trajectories. We also explain how we integrate the ID decodings of the markers along a trajectory to predict the most likely ID for this animal, which can then be used to extract long-term tracks covering the whole lifespan of an individual. See Figure 4 for an overview of our approach.
2.2 Step 1: Linking consecutive detections
The first tracking step considers detections in successive frames. To reduce the number of candidates, we consider only sufficiently close detections (we use approximately 200 pixels, or 12mm).
From these candidate pairs we extract three features:
Euclidean distance between the first detection and its potential successor.
Angular difference of both detections’ orientations on the comb plane.
Manhattan distance between both detections’ ID probabilities.
We use our manually labeled training data to create samples with these features that include both correct and incorrect examples of correspondence. A support vector machine (SVM) with a linear kernel(Cortes and Vapnik, 1995)
is then trained on these samples. We also evaluated the performance of a random forest classifier(Ho, 1995) with comparable results. We use the SVM implemented in the scikit-learn library (Pedregosa et al., 2011a)
. Their implementation of the probability estimate uses Platt’s method(Platt, 1999). This SVM can then be used get the probability of correspondence for pairs of detections that were not included in the training data. To create short tracks (tracklets), we iterate through the recorded data frame by frame and keep a list of open tracklets. Initially, we have one open tracklet for each detection of the first frame. For every time step, we use the SVM to score all new candidates against the last detection of each open tracklet. The Hungarian algorithm is then used to assign the candidate detections to the open tracklets. Tracklets are closed and not further expanded if their best candidate has a probability lower than . Detections that could not be assigned to an existing open tracklet are used to begin a new open tracklet that can be expanded in the next time step.
2.3 Step 2: Merging tracklets
The first step yields a set of short tracklets that do not contain gaps and that could be connected with a high confidence. The second tracking step merges these tracklets into longer tracks that can contain gaps of variable duration (for distributions of tracklet and gap length in our data see Section 3). Note that a tracklet could consist of a single detection or that its corresponding consecutive tracklet could still begin in the next time step without a gap. To reduce computational complexity we define a maximum gap length of 14 time steps ( 4s in our recordings).
Similar to the first tracking step, we use the ground truth dataset to create training samples for a machine learning classifier. We create positive samples (i.e. fragments that should be classified as belonging together) by splitting each manually labeled track once at each time step. Negative samples are generated from each pair of tracks with different IDs which overlapped in time with a maximum gap size of 14. These are also split at all possible time steps. To include both more positive samples and more short track fragments in the training data, we additionally use every correct sub-track of length 3 or less and again split it at all possible locations. This way we generated 1.021.848 training pairs, 7.4% of which were positive samples.
In preliminary tests, we found that for the given task of finding correct correspondences between tracklets, a random forest classifier performed best among a selection of classifiers available in scikit-learn (Boenisch, 2017).
Tracklets with two or more detections allow for more complex and discriminative features compared to those used in the first step. For example, matching tracklets separated by longer gaps may require features that reflect a long-term trend (e.g. the direction of motion).
We implemented 31 different features extractable from tracklet pairs. We then used four different feature selection methods from the scikit-learn library to find the features with the highest predictive power. This evaluation was done by splitting the training data further into a smaller training set and validation set. The methods used were Select-K-Best, Recursive Feature Elimination, Recursive Feature Elimination with Cross-Validation and the Random Forest Feature Importance for all possible feature subset sizes as provided by scikit-learn(Pedregosa et al., 2011b). In all these methods, the same four features (number 1 to 4 in the listing below) performed best according to the ROC AUC score (Spackman, 1989) that proved to be a suitable metric to measure tracking results. Therefore, we chose them as an initial subset.
We then tried to improve the feature subset manually according to more tracking-specific metrics. The metrics we used were the number of tracks in the ground truth validation set that were reconstructed entirely and correctly, and the number of insertions and deletes in the tracks (for further explanation of the metrics see Section 3). We added the features that lead to the highest improvements in these metrics on our validation set. This way, we first added feature 5 and then 6. After adding feature 6, the expansion of the subset with any other feature only lead to a performance decrease in form of more insertions and less complete tracks. We therefore kept the following six features. Visualizations of features 2 to 5 can be found in Figure 5.
Manhattan distance of both tracklets’ bitwise averaged IDs.
Euclidean distance of last detection of tracklet 1 to first detection of tracklet 2.
Forward error: Euclidean distance of linear extrapolation of last motion in first tracklet to first detection in second tracklet.
Backward error: Euclidean distance of linear extrapolation of first motion in second tracklet to last detection in first tracklet.
Angular difference of tag orientation between the last detection of the first tracklet and the first detection of the second tracklet.
Difference of confidence: all IDs in both tracklets are averaged with a bitwise median, we select the bit that is closest to for each tracklet, calculate the absolute difference to (the confidence) and compute the absolute difference of these two confidences.
2.3.1 Track ID assignment
After the second tracking step, we determine the ID of the tracked bee by calculating the median of the bitwise ID probabilities of all detections in the track. The final ID is then determined by binarizing the resulting probabilities for each bit with probability threshold 0.5.
Tracks with a length of several minutes already display a very accurate ID decoding (see Section 3). To calculate longer tracks of up to several days and weeks, we execute the tracking step 1 and step 2 for intervals of one hour and then merge the results to longer tracks based on the assigned ID. This allows us to effectively parallelize the tracking calculation and track the entire season of ten weeks of data in less than a week on a small cluster with less than 100 CPU cores.
3 Results and evaluation
We marked an entire colony of 1953 bees in a two days session and continuously added marked young bees that were bred in an incubation chamber. In total, 2775 bees were marked. The BeesBook system was used to record 10 weeks of continuous image data (3 Hz sample rate) of a one-frame observation hive. The image recordings were stored and processed after the recording season. The computer vision pipeline was executed on a Cray XC30 supercomputer. In total, 3,614,742,669 detections were extracted from 67,972,617 single frames, corresponding to 16,993,154 snapshots of the four cameras. Please note that the data could also be processed in real-time using consumer hardware (Wild et al., 2018).
Two ground truth datasets for the training and evaluation of our method were created manually. A custom program was used to mark the positions of an animal and to define its ID (Mischek, 2016). Details on each dataset can be found in Table 1. To avoid overfitting to specific colony states, the datasets were chosen to contain both high activity (around noon) and low activity (in the early morning hours) periods, different cameras and, therefore, different comb areas. Dataset 2015.1 was used to train and validate classifiers and dataset 2015.2 was used to test their performance.
|Frames||201 (3 fps)||200 (3 fps)|
|False positives||222 (1.23%)||82 (0.75%)|
for testing. The number of detections is the number of tags localized and decoded by the deep learning approach over all frames in the dataset. The number of false positives shows how many times the deep learning pipeline detects a detection when there is none. The number of individuals indicates how many different bees are present in the dataset.
contains 18085 detections from which we extracted 36045 sample pairs (i.e. all pairs with a distance of less than 200 pixels in consecutive frames). These samples were used to train the SVM which is used to link consecutive detections together (tracking step 1). Hyperparameters were determined manually using cross-validation on this dataset. The final model was evaluated on dataset2015.2.
Tracklets for the training and evaluation of a random forest classifier (tracking step 2) were extracted from datasets 2015.1 respectively 2015.2 (see Section 2 for details). Hyperparameters were optimized with hyperopt-sklearn (Komer et al., 2014) on dataset 2015.1 and the optimized model was then tested on dataset 2015.2.
To validate the success of the tracking, we analyzed its impact on several metrics in the tracks, namely:
Proportion of complete tracks
Correctness of resulting tracklets
Length of resulting tracklets
To be able to evaluate the improvement through the presented iterative tracking approach, we compare the results of the two tracking steps to the naive approach of linking the original detections over time based on their initial decoded ID only, in the following referred to as “baseline”. For an overview on the improvements achieved by the different tracking steps see Table 2.
|baseline||after step 1||after step 2||perfect tracking|
|incorrect detection IDs||13.3%||3.9%||1.9%||0.6%|
|incorrect track IDs||63.5||27.2%||18.2%||8.2%|
An important goal of the tracking is to correct IDs of detections which could not be decoded correctly by the computer vision system. Without the tracking algorithm described above, all further behavioral analyses would have to consider this substantial proportion of erroneous decodings. In our dataset, 13.3% of all detections have an incorrectly decoded ID (Wild et al., 2018).
In the ground truth dataset we manually assigned detections that correspond to the same animal to one trajectory. The ground truth data can therefore be considered as the “perfect tracking”. Even on these perfect tracks the median ID assignment algorithm described above provides incorrect IDs for 0.6% of all detections, due to partial occlusions, motion blur and image noise. This represents the lower error bound for the tracking system. As shown in Figure 6, the first tracking step reduces the fraction of incorrect IDs from 13.3% to 3.9% of all detections. The second step further improves this result to only 1.9% incorrect IDs.
Most errors occur in short tracklets (see Figure 7). Therefore, the 1.9% erroneous ID assignments correspond to 18.2% of the resulting tracklets being assigned an incorrect median ID. This is an improvement over the naive baseline and the first tracking step with 63.5% and 27.2% respectively. A perfect tracking could reduce this to 8.2% (see Figure 8).
Proportion of complete tracks
Almost all gaps between detections in our ground truth tracks are no longer than 14 frames (99.76%, see Figure 9). Even though large gaps between detections are rare, long tracks are likely to contain at least one such gap: Only around one third (34.7%) of the ground truth tracks contain no gaps and 77.6% contain only gaps shorter than 14 frames. As displayed in Figure 10, the baseline tracking finds only 10.2% complete tracks without errors (i.e. 30% of all tracks with no gaps). Step 1 is able to correctly assemble 26.5% complete tracks (i.e. around 76.5% of all tracks containing no gaps). Step 2 correctly assembles 70.4% complete tracks (about 90.4% of all tracks with a maximum gap size of less than 14 frames).
Correctness of resulting tracklets
To characterize the type of errors in our tracking results, we define a number of additional metrics. We counted detections that were incorrectly introduced into a track as insertions. Both tracking steps and the baseline inserted only one incorrect detection into another tracklet. Thus less than 1% of both detections and tracklets were affected.
We counted detections that were missing from a tracklet (and were replaced by a gap) as deletions. In the baseline, 32.2% of all detections were missing from their corresponding track (94.6% of all tracks had at least one deletion). After the first step, 1.38% of detections were missing from their track, affecting 26.7% of all tracks. After the second step, 2.37% of all detections and 18.25% of all tracks were still affected.
We also evaluated whether incorrect detections were contained in a track in situations where the correct detection would have been available (instead of a gap) as mismatches, but no resulting tracks contained such mismatches.
Length of resulting tracklets
The ground truth datasets contain only short tracks with a maximum length of one minute. To evaluate the average length of the tracks, we also tracked one hour of data for which no ground truth data was available. The first tracking step yields shorter fragments with an expected length of 2:23 minutes, the second tracking step merges these fragments to tracklets with an expected duration of 6:48 minutes (refer to Figure 11 for tracklet length distributions).
We have presented a multi-step tracking algorithm for fragmentary and partially erroneous detections of honey bee markers. We have applied the proposed algorithm to produce long-term trajectories of all honey bees in a colony of approximately 2000 animals. Our dataset comprises 71 days of continuous positional data at a recording rate of 3 Hz. The presented dataset is by far the most detailed reflection of individual activities of the members of a honey bee colony. The dataset covers the entire lifespan of many hundreds of animals from the day they emerge from their brood cell until the day they die. Honey bees rely on a flexible but generally age-dependent division of labor. Hence, our dataset reflects all essential aspects of a self-sustaining colony, from an egg-laying queen and brood rearing young workers, to food collection, and colony defense. We have released a three days sample dataset for the interested reader (Boenisch et al., 2018). Our implementation of the proposed tracking algorithm is available online111https://github.com/BioroboticsLab/bb_tracking.
The tracking framework presented in the previous sections is an essential part of the BeesBook system. It provides a computationally efficient approach to determine the correct IDs for more than 98% of the individuals in the honey bee hive without using extra bits for error correction.
Although it is possible to use error correction with 12 bit markers, this would reduce the number of coding bits and therefore the number of observable animals. While others chose to increase the number of bits on the marker, we solved the problem in the tracking stage. With the proposed system, we were able to reduce hardware costs for cameras and storage. When applied to the raw output of the image decoding step, the accuracy of other systems that use error-correction (for example Mersch et al. (2013)) may even be improved further.
Our system provides highly accurate movement paths of bees. Given a long-term observation of several weeks, these paths, however, can still be considered short fragments. Since the IDs of these tracklets are very accurate, they can now be linked by matching IDs only.
Still, some aspects of the system can be improved. To train our classifiers, we need a sufficiently large, manually labeled dataset. Rice et al. (2015) proposed a method to create a similar dataset interactively, reducing the required manual work. Also, the circular coding scheme of our markers causes some bit configurations to appear similar under certain object poses. This knowledge could be integrated into our ID determination algorithm. The IDs along a trajectory might not provide an equal amount of information. Some might be recorded under fast motion and are therefore less reliable. Other detections could have been recorded from a still bee whose tag was partially occluded. Considering similar readings as less informative might improve the ID accuracy of our method. Still, with the proposed method there are only 1.9% detections incorrectly decoded, mostly in very short tracklets.
The resulting trajectories can now be used for further analyses of individual honey bee behavior or interactions in the social network. In addition to the three day dataset published alongside this paper, we plan to publish two more datasets covering more than 60 days of recordings, each. With this data we can investigate how bees acquire information in the colony and how that experience modulates future behavior and interactions. We hope that through this work we can interest researchers to join the collective effort of investigating the individual and collective intelligence of the honey bee, a model organism that bears a vast number of fascinating research questions.
Conflict of Interest Statement
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
FW received funding from the German Academic Exchange Service (DAAD). DD received funding from the Andrea von Braun Foundation. This work was in part funded by the Klaus Tschira Foundation. We also acknowledge the support by the Open Access Publication Initiative of the Freie Universität Berlin.
We are indebted to the help of Jakob Mischek for his preliminary work and his help with creating the ground truth data.
- Balbuena et al. (2012) M. S. Balbuena, J. Molinas, and W. M. Farina. Honeybee recruitment to scented food sources: correlations between in-hive social interactions and foraging decisions. Behavioral Ecology and Sociobiology, 66(3):445–452, 2012. ISSN 1432-0762. doi: 10.1007/s00265-011-1290-3. URL http://dx.doi.org/10.1007/s00265-011-1290-3. 00011.
- Baracchi and Cini (2014) David Baracchi and Alessandro Cini. A Socio-Spatial Combined Approach Confirms a Highly Compartmentalised Structure in Honeybees. Ethology, 120(12):1167–1176, December 2014. ISSN 1439-0310. doi: 10.1111/eth.12290. URL http://onlinelibrary.wiley.com/doi/10.1111/eth.12290/abstract.
- Berman et al. (2014) Gordon J. Berman, Daniel M. Choi, William Bialek, and Joshua W. Shaevitz. Mapping the stereotyped behaviour of freely moving fruit flies. Journal of The Royal Society Interface, 11(99):20140672, October 2014. ISSN 1742-5689, 1742-5662. doi: 10.1098/rsif.2014.0672. URL http://rsif.royalsocietypublishing.org/content/11/99/20140672.
- Betke and Wu (2016) Margrit Betke and Zheng Wu. Data association for multi-object visual tracking. Synthesis Lectures on Computer Vision, 6(2):1–120, 2016. doi: http://dx.doi.org/10.2200/S00713ED1V01Y201603COV009.
- Biesmeijer and Seeley (2005) Jacobus C. Biesmeijer and Thomas D. Seeley. The use of waggle dance information by honey bees throughout their foraging careers. Behavioral Ecology and Sociobiology, 59(1):133–142, 2005. ISSN 1432-0762. doi: 10.1007/s00265-005-0019-6. URL http://dx.doi.org/10.1007/s00265-005-0019-6. 00111.
- Blut et al. (2017) Christina Blut, Alessandro Crespi, Danielle Mersch, Laurent Keller, Linlin Zhao, Markus Kollmann, Benjamin Schellscheidt, Carsten Fülber, and Martin Beye. Automated computer-based detection of encounter behaviours in groups of honeybees. Scientific Reports, 7(1):17663, December 2017. ISSN 2045-2322. doi: 10.1038/s41598-017-17863-4. URL https://www.nature.com/articles/s41598-017-17863-4. 00000.
- Boenisch (2017) Franziska Boenisch. Feature Engineering and Probabilistic Tracking on Honey Bee Trajectories. Bachelor Thesis, 2017.
- Boenisch et al. (2018) Franziska Boenisch, Benjamin Rosemann, Benjamin Wild, Fernando Wario, David Dormagen, and Tim Landgraf. BeesBook Recording Season 2015 Sample, 2018. synapse.org/#!Synapse:syn11737848 DOI: 10.7303/syn11737848.1.
- Brian R. Johnson (2010) Brian R. Johnson. Division of labor in honeybees: form, function, and proximate mechanisms. Behavioral Ecology and Sociobiology, 64(3):305–316, January 2010. ISSN 0340-5443. doi: 10.1007/s00265-009-0874-7. URL http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2810364/.
- Cortes and Vapnik (1995) Corinna Cortes and Vladimir Vapnik. Support-Vector Networks. 20(3):273–297, 1995. ISSN 1573-0565. doi: 10.1023/A:1022627411411.
- Couvillon et al. (2014) Margaret J. Couvillon, Roger Schürch, and Francis L. W. Ratnieks. Waggle Dance Distances as Integrative Indicators of Seasonal Foraging Challenges. PLOS ONE, 9(4):e93495, February 2014. ISSN 1932-6203. doi: 10.1371/journal.pone.0093495. URL http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0093495.
- Cox (1993) Ingemar J. Cox. A review of statistical data association techniques for motion correspondence. International Journal of Computer Vision, 10(1):53–66, February 1993. ISSN 0920-5691, 1573-1405. doi: 10.1007/BF01440847. URL https://link.springer.com/article/10.1007/BF01440847.
- Crall et al. (2015) James D. Crall, Nick Gravish, Andrew M. Mountcastle, and Stacey A. Combes. BEEtag: A Low-Cost, Image-Based Tracking System for the Study of Animal Behavior and Locomotion. 10(9):e0136487, 2015. ISSN 1932-6203. doi: 10.1371/journal.pone.0136487.
- De Marco and Farina (2001) Rodrigo De Marco and Walter Farina. Changes in food source profitability affect the trophallactic and dance behavior of forager honeybees (Apis mellifera L.). Behavioral Ecology and Sociobiology, 50(5):441–449, October 2001. ISSN 0340-5443, 1432-0762. doi: 10.1007/s002650100382. URL https://link.springer.com/article/10.1007/s002650100382.
- Dell et al. (2014a) Anthony I. Dell, John A. Bender, Kristin Branson, Iain D. Couzin, Gonzalo G. de Polavieja, Lucas P.J.J. Noldus, Alfonso Pérez-Escudero, Pietro Perona, Andrew D. Straw, Martin Wikelski, and Ulrich Brose. Automated image-based tracking and its application in ecology. Trends in Ecology & Evolution, 29(7):417–428, July 2014a. ISSN 01695347. doi: 10.1016/j.tree.2014.05.004. URL http://linkinghub.elsevier.com/retrieve/pii/S0169534714001074. 00099.
- Dell et al. (2014b) Anthony I. Dell, John A. Bender, Kristin Branson, Iain D. Couzin, Gonzalo G. de Polavieja, Lucas P.J.J. Noldus, Alfonso Pérez-Escudero, Pietro Perona, Andrew D. Straw, Martin Wikelski, and Ulrich Brose. Automated image-based tracking and its application in ecology. Trends in Ecology & Evolution, 29(7):417–428, July 2014b. ISSN 01695347. doi: 10.1016/j.tree.2014.05.004. URL http://linkinghub.elsevier.com/retrieve/pii/S0169534714001074. 00099.
- Eyjolfsdottir et al. (2016) Eyrun Eyjolfsdottir, Kristin Branson, Yisong Yue, and Pietro Perona. Learning recurrent representations for hierarchical behavior modeling. arXiv:1611.00094 [cs], October 2016. URL http://arxiv.org/abs/1611.00094. arXiv: 1611.00094.
- Fasciano et al. (2013) Thomas Fasciano, Hoan Nguyen, Anna Dornhaus, and Min C. Shin. Tracking multiple ants in a colony. In 2013 IEEE Workshop on Applications of Computer Vision (WACV), pages 534–540. IEEE, jan 2013. ISBN 978-1-4673-5054-9. doi: 10.1109/WACV.2013.6475065. URL http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=6475065.
- Gernat et al. (2018) Tim Gernat, Vikyath D. Rao, Martin Middendorf, Harry Dankowicz, Nigel Goldenfeld, and Gene E. Robinson. Automated monitoring of behavior reveals bursty interaction patterns and rapid spreading dynamics in honeybee social networks. Proceedings of the National Academy of Sciences, 115(7):1433–1438, February 2018. ISSN 0027-8424, 1091-6490. doi: 10.1073/pnas.1713568115. URL http://www.pnas.org/lookup/doi/10.1073/pnas.1713568115.
- Goyret and Farina (2005) Joaquín Goyret and Walter M. Farina. Non-random nectar unloading interactions between foragers and their receivers in the honeybee hive. Naturwissenschaften, 92(9):440–443, 2005. ISSN 1432-1904. doi: 10.1007/s00114-005-0016-7. URL http://dx.doi.org/10.1007/s00114-005-0016-7. 00027.
- Grüter and Farina (2009) Christoph Grüter and Walter M. Farina. Past Experiences Affect Interaction Patterns Among Foragers and Hive-Mates in Honeybees. Ethology, 115(8):790–797, August 2009. ISSN 1439-0310. doi: 10.1111/j.1439-0310.2009.01670.x. URL http://onlinelibrary.wiley.com/doi/10.1111/j.1439-0310.2009.01670.x/abstract. 00017.
- Grüter and Ratnieks (2011) Christoph Grüter and Francis L. W. Ratnieks. Honeybee foragers increase the use of waggle dance information when private information becomes unrewarding. Animal Behaviour, 81(5):949–954, 2011. ISSN 00033472. doi: 10.1016/j.anbehav.2011.01.014. URL http://dx.doi.org/10.1016/j.anbehav.2011.01.014.
- Grüter et al. (2006) Christoph Grüter, Luis E Acosta, and Walter M Farina. Propagation of olfactory information within the honeybee hive. Behavioral Ecology and Sociobiology, 60(5):707–715, 2006. ISSN 0340-5443.
- Ho (1995) Tin Kam Ho. Random decision forests. In Document Analysis and Recognition, 1995., Proceedings of the Third International Conference On, volume 1, pages 278–282. IEEE, 1995.
- Huang et al. (2008) Chang Huang, Bo Wu, and Ramakant Nevatia. Robust object tracking by hierarchical association of detection responses. In European Conference on Computer Vision, pages 788–801. Springer, 2008.
- Kabra et al. (2013) Mayank Kabra, Alice A. Robie, Marta Rivera-Alba, Steven Branson, and Kristin Branson. JAABA: interactive machine learning for automatic annotation of animal behavior. Nature Methods, 10(1):64–67, January 2013. ISSN 1548-7105. doi: 10.1038/nmeth.2281. URL https://www.nature.com/articles/nmeth.2281.
- Komer et al. (2014) Brent Komer, James Bergstra, and Chris Eliasmith. Hyperopt-sklearn: automatic hyperparameter configuration for scikit-learn. In ICML workshop on AutoML, 2014.
- Krause et al. (2013) Jens Krause, Stefan Krause, Robert Arlinghaus, Ioannis Psorakis, Stephen Roberts, and Christian Rutz. Reality mining of animal social systems. Trends in Ecology & Evolution, 28(9):541–551, September 2013. ISSN 01695347. doi: 10.1016/j.tree.2013.06.002. URL http://linkinghub.elsevier.com/retrieve/pii/S0169534713001468.
- Kuhn (1955) Harold W. Kuhn. The Hungarian method for the assignment problem. 2:83–97, 1955.
- Kühl and Burghardt (2013) Hjalmar S. Kühl and Tilo Burghardt. Animal biometrics: quantifying and detecting phenotypic appearance. Trends in Ecology & Evolution, 28(7):432–441, July 2013. ISSN 01695347. doi: 10.1016/j.tree.2013.02.013. URL http://linkinghub.elsevier.com/retrieve/pii/S0169534713000657.
- Luo et al. (2014) Wenhan Luo, Junliang Xing, Anton Milan, Xiaoqin Zhang, Wei Liu, Xiaowei Zhao, and Tae-Kyun Kim. Multiple Object Tracking: A Literature Review. arXiv:1409.7618 [cs], September 2014. URL http://arxiv.org/abs/1409.7618. 00031 arXiv: 1409.7618.
- Mersch et al. (2013) Danielle P. Mersch, Alessandro Crespi, and Laurent Keller. Tracking individuals shows spatial fidelity is a key regulator of ant social organization. Science, 340(6136):1090–3, 2013. ISSN 1095-9203. doi: 10.1126/science.1234316. URL http://www.sciencemag.org/cgi/doi/10.1126/science.1234316http://www.ncbi.nlm.nih.gov/pubmed/23599264.
- Mischek (2016) Jakob Mischek. Probabilistisches Tracking von Bienenpfaden. Master Thesis, Freie Universität Berlin, September 2016.
- Naug (2008) Dhruba Naug. Structure of the social network and its influence on transmission dynamics in a honeybee colony. Behavioral Ecology and Sociobiology, 62(11):1719–1725, September 2008. ISSN 0340-5443, 1432-0762. doi: 10.1007/s00265-008-0600-x. URL http://link.springer.com/10.1007/s00265-008-0600-x.
- Pedregosa et al. (2011a) Fabian Pedregosa, Gaël Varoquaux, Alexandre Gramfort, Vincent Michel, Bertrand Thirion, Olivier Grisel, Mathieu Blondel, Peter Prettenhofer, Ron Weiss, Vincent Dubourg, Jake Vanderplas, Alexandre Passos, David Cournapeau, Matthieu Brucher, Matthieu Perrot, and Édouard Duchesnay. Scikit-learn: Machine Learning in Python. Journal of Machine Learning Research, 12:2825–2830, October 2011a. URL http://jmlr.csail.mit.edu/papers/v12/pedregosa11a.html.
- Pedregosa et al. (2011b) Fabian Pedregosa, Gaël Varoquaux, Alexandre Gramfort, Vincent Michel, Bertrand Thirion, Olivier Grisel, Mathieu Blondel, Peter Prettenhofer, Ron Weiss, Vincent Dubourg, Jake Vanderplas, Alexandre Passos, David Cournapeau, Matthieu Brucher, Matthieu Perrot, and Édouard Duchesnay. Scikit-learn: Machine Learning in Python. Journal of Machine Learning Research, 12:2825–2830, October 2011b. URL http://jmlr.csail.mit.edu/papers/v12/pedregosa11a.html. 07769.
- Platt (1999) John C. Platt. Probabilistic Outputs for Support Vector Machines and Comparisons to Regularized Likelihood Methods. In Advances in Large Margin Classifiers, pages 61–74. MIT Press, 1999.
- Pérez-Escudero et al. (2014) Alfonso Pérez-Escudero, Julián Vicente-Page, Robert C Hinz, Sara Arganda, and Gonzalo G de Polavieja. idTracker: tracking individuals in a group by automatic identification of unmarked animals. Nature Methods, 11(7):743–748, July 2014. ISSN 1548-7091, 1548-7105. doi: 10.1038/nmeth.2994. URL http://www.nature.com/articles/nmeth.2994.
- Rice et al. (2015) Lance Rice, Anna Dornhausy, and Min C. Shin. Efficient Training of Multiple Ant Tracking. pages 117–123. IEEE, January 2015. ISBN 978-1-4799-6683-7. doi: 10.1109/WACV.2015.23. URL http://ieeexplore.ieee.org/document/7045877/. 00002.
- Richter and Waddington (1993) Monica Raveret Richter and Keith D. Waddington. Past foraging experience influences honey bee dance behaviour. Animal Behaviour, 46(1):123–128, July 1993. ISSN 0003-3472. doi: 10.1006/anbe.1993.1167. URL http://www.sciencedirect.com/science/article/pii/S000334728371167X. 00060.
- Robinson (1992) Gene E. Robinson. Regulation of division of labor in insect societies. Annual review of entomology, 37(1):637–665, 1992. URL http://www.annualreviews.org/doi/pdf/10.1146/annurev.en.37.010192.003225.
- Rosemann. (2017) Benjamin Rosemann. Ein erweiterbares system für experimente mit multi-target tracking von markierten bienen. Master Thesis, Freie Universität Berlin, 2017.
- Seeley (2010) Thomas D. Seeley. Honeybee Democracy. Princeton University Press, Princeton, New Jersey, 2010. ISBN 978-0-691-14721-5. URL http://www.amazon.com/Honeybee-Democracy-Thomas-D-Seeley/dp/0691147213. 00399.
- Spackman (1989) Kent A. Spackman. Signal detection theory: Valuable tools for evaluating inductive learning. In Proceedings of the Sixth International Workshop on Machine Learning, pages 160–163. Morgan Kaufmann Publishers Inc., 1989. ISBN 978-1-55860-036-2. URL http://dl.acm.org/citation.cfm?id=102118.102172.
- Thompson (1983) Thomas M Thompson. From error-correcting codes through sphere packings to simple groups. Number 21. Cambridge University Press, 1983.
- von Frisch (1965) Karl von Frisch. Tanzsprache und Orientierung der Bienen. Springer, 1965. ISBN 978-3-642-94916-6. URL http://www.springer.com/us/book/9783642949173.
- Wang et al. (2014) L. Wang, N. H. C. Yung, and L. Xu. Multiple-Human Tracking by Iterative Data Association and Detection Update. IEEE Transactions on Intelligent Transportation Systems, 15(5):1886–1899, October 2014. ISSN 1524-9050. doi: 10.1109/TITS.2014.2303196.
- Wang and Yeung (2013) Naiyan Wang and Dit-Yan Yeung. Learning a Deep Compact Image Representation for Visual Tracking. In Proceedings of the 26th International Conference on Neural Information Processing Systems - Volume 1, NIPS’13, pages 809–817, USA, 2013. Curran Associates Inc. URL http://dl.acm.org/citation.cfm?id=2999611.2999702.
- Wario (2017) Fernando Wario. A computer vision based system for the automatic analysis of social networks in honey bee colonies. PhD dissertation, Freie Universität Berlin, 2017.
- Wario et al. (2015) Fernando Wario, Benjamin Wild, Margaret Jane Couvillon, Raúl Rojas, and Tim Landgraf. Automatic methods for long-term tracking and the detection and decoding of communication dances in honeybees. Behavioral and Evolutionary Ecology, page 103, 2015. doi: 10.3389/fevo.2015.00103. URL http://journal.frontiersin.org/article/10.3389/fevo.2015.00103/full.
- Wario et al. (2017) Fernando Wario, Benjamin Wild, Raúl Rojas, and Tim Landgraf. Automatic detection and decoding of honey bee waggle dances. arXiv preprint arXiv:1708.06590, 2017.
- Wild et al. (2018) Benjamin Wild, Leon Sixt, and Tim Landgraf. Automatic localization and decoding of honeybee markers using deep convolutional neural networks. arXiv:1802.04557 [cs], February 2018. URL http://arxiv.org/abs/1802.04557. arXiv: 1802.04557.
- Wiltschko et al. (2015) AlexanderB. Wiltschko, MatthewJ. Johnson, Giuliano Iurilli, RalphE. Peterson, JesseM. Katon, StanL. Pashkovski, VictoriaE. Abraira, RyanP. Adams, and SandeepRobert Datta. Mapping Sub-Second Structure in Mouse Behavior. Neuron, 88(6):1121–1135, December 2015. ISSN 0896-6273. doi: 10.1016/j.neuron.2015.11.031. URL http://www.sciencedirect.com/science/article/pii/S0896627315010375.
- Wu and Nevatia (2007) Bo Wu and Ram Nevatia. Detection and Tracking of Multiple, Partially Occluded Humans by Bayesian Combination of Edgelet based Part Detectors. International Journal of Computer Vision, 75(2):247–266, November 2007. ISSN 0920-5691, 1573-1405. doi: 10.1007/s11263-006-0027-7. URL https://link.springer.com/article/10.1007/s11263-006-0027-7.
Wu et al. (2013)
Y. Wu, J. Lim, and M. H. Yang.
Online Object Tracking: A Benchmark.
2013 IEEE Conference on Computer Vision and Pattern Recognition, pages 2411–2418, June 2013. doi: 10.1109/CVPR.2013.312.