Multi-Lane Capsule Network for Classifying Images With Complex Background

Abstract

Capsule Network (CapsNet) is a novel structure for deep neural network, mapping the region of target instance to vectors and matrices rather than scalars. This process is enabled by dynamic routing algorithm which help CapsNet to achieve more robust capacity with fewer parameters than traditional CNNs. However, one drawback of Capsule is that it turns to account for everything in the image, which leads to a poor performance when the backgrounds are too varied to model with a reasonable sized net. We proposed a multi-lane capsule network with strict-squash (MLSCN) to solve this problem. In MLSCN, we designed a novel Capsule based network structure, replaced the Squash function and optimized the implementation of dropout. To validate modification proposed in the paper, we conducted extensive experiments on three public datasets which include MNIST, affNIST and CIFAR10. Ablation experiments were also conducted to analyze contributions of each component of MLSCN. Experimental results show that MLSCN outperforms the original CapsNet in multiple benchmarks. Our model boosts the classification accuracy of CIFAR10 about 17% with negligible parameter increase compare to original CapsNet. Besides, the proposed model achieves an accuracy of 65.37%, while original CapsNet is only 41.62% on MNIST-BC. INDEX TERMS Capsule network, multi-lane, strict-squash, drop-circuit, lane-filter, complex background.

CTSF: An End-to-End Efficient Neural Network for Chinese Text with Skeleton Feature

“…We added a vertical anchor regression mechanism to CTPN to detect text lines, and at the same time use PVANet's image feature extraction module to optimize and accelerate the detection network [12][13]. Finally, this paper also proposes a literacy model based on spatial pyramid pooling and fusion of Chinese character skeleton features, which reduces the loss of image information caused by fixed input image size through spatial pyramid pooling; improves the accuracy of Chinese character recognition by adding Chinese character skeleton features [14][15][16].…”

Towards Knowledge Enhanced Language Model for Machine Reading Comprehension

“…When related entities are mentioned in the article, the semantic vector is used as external knowledge for direct accumulation, leading to deficiencies in both knowledge extraction and knowledge fusion. To take full advantage of facts in KGs which contains rich language patterns, we propose a novel module CapsBERT as a preprocessing step, which fine-tunes pre-trained BERT [36] and utilizes a capsule neural network (CapsNet) [37] for knowledge graph representation. BERT is a state-of-the-art contextual language representation model built on a multilayer bidirectional Transformer encoder, which is able to capture complex linguistic phenomena.…”

Feature Fusion-Based Improved Capsule Network for sEMG Signal Recognition

“…Equation Squash [14] v j � (‖s j ‖ 2 /(1 + ‖s j ‖ 2 )(s j /‖s j ‖)) HSquash [22] v j � ‖s j /4‖ 2 /1 + ‖s j /4‖ 2 s j /‖s j ‖ Strict-squash [23] v j � 0.69‖s j ‖ 2 * 2 − 0.6‖s j ‖− 1.115 Squash-4 [24] v j � ‖s j ‖ 2 /0.5 + ‖s j ‖ 2 s j /‖s j ‖ e-Squash…”