mlsanityData to 7 Oct 2026

Capsule Network is Not More Robust than Convolutional Network

arXiv 2103.15459 ↗First posted 29 Mar 2021cs.CVcited by 12

Not on the map. Why?

Abstract (arXiv)

The Capsule Network is widely believed to be more robust than Convolutional Networks. However, there are no comprehensive comparisons between these two networks, and it is also unknown which components in the CapsNet affect its robustness. In this paper, we first carefully examine the special designs in CapsNet that differ from that of a ConvNet commonly used for image classification. The examination reveals five major new/different components in CapsNet: a transformation process, a dynamic routing layer, a squashing function, a marginal loss other than cross-entropy loss, and an additional class-conditional reconstruction loss for regularization. Along with these major differences, we conduct comprehensive ablation studies on three kinds of robustness, including affine transformation, overlapping digits, and semantic representation. The study reveals that some designs, which are thought critical to CapsNet, actually can harm its robustness, i.e., the dynamic routing layer and the transformation process, while others are beneficial for the robustness. Based on these findings, we propose enhanced ConvNets simply by introducing the essential components behind the CapsNet's success. The proposed simple ConvNets can achieve better robustness than the CapsNet.

Key facts. Cited by 12 papers in the map (data through 7 Oct 2026). Builds on 18 of its 26 references that are in the map. First posted on arXiv 29 Mar 2021.

Builds on

18 of 26 references.

References · 18

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2016Densely Connected Convolutional Networkscited by 6,691
  3. 2017Dynamic Routing Between Capsulescited by 689
  4. 2017Relation Networks for Object Detectioncited by 305
  5. 2019Local Relation Networks for Image Recognitioncited by 210
  6. 2019Stacked Capsule Autoencoderscited by 82
  7. 2019DeepCaps: Going Deeper with Capsule Networkscited by 42
  8. 2021Efficient-CapsNet: Capsule Network with Self-Attention Routingcited by 41
  9. 2019Capsule Routing via Variational Bayescited by 32
  10. 2018Group Equivariant Capsule Networkscited by 31
  11. 2019Detecting and Diagnosing Adversarial Images with Class-Conditional Capsule Reconstructionscited by 30
  12. 2020Capsules with Inverted Dot-Product Attention Routingcited by 25
6 more
  1. 2019Improving the Robustness of Capsule Networks to Image Affine Transformationscited by 21
  2. 2021Effective and Efficient Vote Attack on Capsule Networkscited by 14
  3. 2020Interpretable Graph Capsule Networks for Object Recognitioncited by 12
  4. 2019On the Vulnerability of Capsule Networks to Adversarial Attackscited by 11
  5. 2018Sparse Unsupervised Capsules Generalize Bettercited by 8
  6. 2020Deflecting Adversarial Attackscited by 7

Built on by

Cited by year

Citing papers per year: 2021: 1; 2022: 4; 2023: 5; 2024: 1; 2025: 1.

12 papers.

  1. 2022LAS-AT: Adversarial Training with Learnable Attack Strategycited by 94
  2. 2022Prior-Guided Adversarial Initialization for Fast Adversarial Trainingcited by 34
  3. 2023Benchmarking Robustness of Adaptation Methods on Pre-trained Vision-Language Modelscited by 30
  4. 2022Recognizing Object by Components with Human Prior Knowledge Enhances Adversarial Robustness of Deep Neural Networkscited by 18
  5. 2023Reliable Evaluation of Adversarial Transferabilitycited by 12
  6. 20223D-UCaps: 3D Capsules Unet for Volumetric Image Segmentationcited by 7
  7. 2023Towards Robust Prompts on Vision-Language Modelscited by 7
  8. 2023Why Capsule Neural Networks Do Not Scale: Challenging the Dynamic Parse-Tree Assumptioncited by 6
  9. 2021Recurrent Attention Models with Object-centric Capsule Representation for Multi-object Recognitioncited by 2
  10. 2023Explainability and Robustness of Deep Visual Classification Modelscited by 0
  11. 2024CardioCaps: Attention-based Capsule Network for Class-Imbalanced Echocardiogram Classificationcited by 0
  12. 2025Explainable AI-Guided Efficient Approximate DNN Generation for Multi-Pod Systolic Arrayscited by 0