Trace-back Along Capsules and Its Application on Semantic Segmentation

by   Tao Sun, et al.

In this paper, we propose a capsule-based neural network model to solve the semantic segmentation problem. By taking advantage of the extractable part-whole dependencies available in capsule layers, we derive the probabilities of the class labels for individual capsules through a recursive, layer-by-layer procedure. We model this procedure as a traceback pipeline and take it as a central piece to build an end-to-end segmentation network. Under the proposed framework, image-level class labels and object boundaries are jointly sought in an explicit manner, which poses a significant advantage over the state-of-the-art fully convolutional network (FCN) solutions. Experiments conducted on modified MNIST and neuroimages demonstrate that our model considerably enhance the segmentation performance compared to the leading FCN variant.


page 6

page 10


Object Boundary Guided Semantic Segmentation

Semantic segmentation is critical to image content understanding and obj...

Graph-FCN for image semantic segmentation

Semantic segmentation with deep learning has achieved great progress in ...

Residual Pyramid FCN for Robust Follicle Segmentation

In this paper, we propose a pyramid network structure to improve the FCN...

Context Encoding for Semantic Segmentation

Recent work has made significant progress in improving spatial resolutio...

Deep Variational Instance Segmentation

Instance Segmentation, which seeks to obtain both class and instance lab...

Dilated Temporal Fully-Convolutional Network for Semantic Segmentation of Motion Capture Data

Semantic segmentation of motion capture sequences plays a key part in ma...

Joint Stem Detection and Crop-Weed Classification for Plant-specific Treatment in Precision Farming

Applying agrochemicals is the default procedure for conventional weed co...