Capsules for Object Segmentation

Abstract

Convolutional neural networks (CNNs) have shown remarkable results over the last several years for a wide range of computer vision tasks. A new architecture recently introduced by Sabour et al., referred to as a capsule networks with dynamic routing, has shown great initial results for digit recognition and small image classification. The success of capsule networks lies in their ability to preserve more information about the input by replacing max-pooling layers with convolutional strides and dynamic routing, allowing for preservation of part-whole relationships in the data. This preservation of the input is demonstrated by reconstructing the input from the output capsule vectors. Our work expands the use of capsule networks to the task of object segmentation for the first time in the literature. We extend the idea of convolutional capsules with locally-connected routing and propose the concept of deconvolutional capsules. Further, we extend the masked reconstruction to reconstruct the positive input class. The proposed convolutional-deconvolutional capsule network, called SegCaps, shows strong results for the task of object segmentation with substantial decrease in parameter space. As an example application, we applied the proposed SegCaps to segment pathological lungs from low dose CT scans and compared its accuracy and efficiency with other U-Net-based architectures. SegCaps is able to handle large image sizes (512 x 512) as opposed to baseline capsules (typically less than 32 x 32). The proposed SegCaps reduced the number of parameters of U-Net architecture by 95.4% while still providing a better segmentation accuracy.

References in corpus (1)

• Adam: A Method for Stochastic Optimization

Cited by in corpus (36)

• Efficient-CapsNet: Capsule Network with Self-Attention Routing
• Blockchain-Federated-Learning and Deep Learning Models for COVID-19 detection using CT Imaging
• Multi-Interest Network with Dynamic Routing for Recommendation at Tmall
• Point2SpatialCapsule: Aggregating Features and Spatial Relationships of Local Regions on Point Clouds using Spatial-aware Capsules
• 1D-CapsNet-LSTM: A Deep Learning-Based Model for Multi-Step Stock Index Forecasting
• Capsule Networks Need an Improved Routing Algorithm
• 1D-Convolutional Capsule Network for Hyperspectral Image Classification
• Unsupervised part representation by Flow Capsules
• Domain specific cues improve robustness of deep learning based segmentation of ct volumes
• Siamese Capsule Networks
• DESCNet: Developing Efficient Scratchpad Memories for Capsule Network Hardware
• Dynamic Layer Aggregation for Neural Machine Translation with Routing-by-Agreement
• Quantum Feature-Empowered Deep Classification for Fast Mangrove Mapping
• Information Aggregation for Multi-Head Attention with Routing-by-Agreement
• U-Det: A Modified U-Net architecture with bidirectional feature network for lung nodule segmentation
• Capsule networks with non-iterative cluster routing
• Blockchain based Privacy-Preserved Federated Learning for Medical Images: A Case Study of COVID-19 CT Scans
• Building Deep, Equivariant Capsule Networks
• Multi-modal Capsule Routing for Actor and Action Video Segmentation Conditioned on Natural Language Queries
• Multi-organ Segmentation over Partially Labeled Datasets with Multi-scale Feature Abstraction
• LGAN: Lung Segmentation in CT Scans Using Generative Adversarial Network
• Examining the Benefits of Capsule Neural Networks
• TCDCaps: Visual Tracking via Cascaded Dense Capsules
• Improving the Robustness of Capsule Networks to Image Affine Transformations
• Recognizing Image Objects by Relational Analysis Using Heterogeneous Superpixels and Deep Convolutional Features
• Dynamic Routing on Deep Neural Network for Thoracic Disease Classification and Sensitive Area Localization
• Capsule network with shortcut routing
• Detecting Adversarial Patches with Class Conditional Reconstruction Networks
• Crosslink-Net: Double-branch Encoder Segmentation Network via Fusing Vertical and Horizontal Convolutions
• ICE-GAN: Identity-aware and Capsule-Enhanced GAN with Graph-based Reasoning for Micro-Expression Recognition and Synthesis
• TraceCaps: A Capsule-based Neural Network for Semantic Segmentation
• Leveraging Domain Adaptation for Low-Resource Geospatial Machine Learning
• Sequential Routing Framework: Fully Capsule Network-based Speech Recognition
• Crossover-Net: Leveraging the Vertical-Horizontal Crossover Relation for Robust Segmentation
• Automatically Designing CNN Architectures for Medical Image Segmentation
• An Efficient Agreement Mechanism in CapsNets By Pairwise Product