CVJul 30, 2020

An Improvement for Capsule Networks using Depthwise Separable Convolution

arXiv:2007.15167v24 citations
AI Analysis

This work addresses a critical issue in computer vision for Capsule Networks, offering an incremental improvement by integrating a known technique to enhance efficiency and performance.

The paper tackles the problem of Capsule Networks' performance being challenged by image backgrounds by proposing an improved architecture that replaces Standard Convolution with Depthwise Separable Convolution, resulting in reduced parameters, increased stability, and competitive accuracy, with the model outperforming standard ones on 64x64 pixel images.

Capsule Networks face a critical problem in computer vision in the sense that the image background can challenge its performance, although they learn very well on training data. In this work, we propose to improve Capsule Networks' architecture by replacing the Standard Convolution with a Depthwise Separable Convolution. This new design significantly reduces the model's total parameters while increases stability and offers competitive accuracy. In addition, the proposed model on $64\times64$ pixel images outperforms standard models on $32\times32$ and $64\times64$ pixel images. Moreover, we empirically evaluate these models with Deep Learning architectures using state-of-the-art Transfer Learning networks such as Inception V3 and MobileNet V1. The results show that Capsule Networks can perform comparably against Deep Learning models. To the best of our knowledge, we believe that this is the first work on the integration of Depthwise Separable Convolution into Capsule Networks.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes