mlsanityData to 14 Oct 2026

Recurrent Attention Models with Object-centric Capsule Representation for Multi-object Recognition

arXiv 2110.04954 ↗First posted 11 Oct 2021cs.CV cs.LG q-bio.NCcited by 2

Not on the map. Why?

Abstract (arXiv)

The visual system processes a scene using a sequence of selective glimpses, each driven by spatial and object-based attention. These glimpses reflect what is relevant to the ongoing task and are selected through recurrent processing and recognition of the objects in the scene. In contrast, most models treat attention selection and recognition as separate stages in a feedforward process. Here we show that using capsule networks to create an object-centric hidden representation in an encoder-decoder model with iterative glimpse attention yields effective integration of attention and recognition. We evaluate our model on three multi-object recognition tasks; highly overlapping digits, digits among distracting clutter and house numbers, and show that it learns to effectively move its glimpse window, recognize and reconstruct the objects, all with only the classification as supervision. Our work takes a step toward a general architecture for how to integrate recurrent object-centric representation into the planning of attentional glimpses.

Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Builds on 39 of its 68 references that are in the map. First posted on arXiv 11 Oct 2021.

Builds on

39 of 68 references.

References · 39

  1. 2017Attention Is All You Needcited by 60,361
  2. 2014Adam: A Method for Stochastic Optimizationcited by 37,240
  3. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,305
  4. 2017Squeeze-and-Excitation Networkscited by 4,789
  5. 2015Spatial Transformer Networkscited by 1,429
  6. 2015Show, Attend and Tell: Neural Image Caption Generation with Visual Attentioncited by 1,423
  7. 2021Perceiver: General Perception with Iterative Attentioncited by 1,104
  8. 2020Object-Centric Learning with Slot Attentioncited by 876
  9. 2017Dynamic Routing Between Capsulescited by 689
  10. 2019MONet: Unsupervised Scene Decomposition and Representationcited by 392
  11. 2017Residual Attention Network for Image Classificationcited by 386
  12. 2019Multi-Object Representation Learning with Iterative Variational Inferencecited by 360
27 more
  1. 2014Recurrent Models of Visual Attentioncited by 349
  2. 2016SCA-CNN: Spatial and Channel-wise Attention in Convolutional Networks for Image Captioningcited by 219
  3. 2019Recurrent Independent Mechanismscited by 219
  4. 2016Attend, Infer, Repeat: Fast Scene Understanding with Generative Modelscited by 209
  5. 2015DRAW: A Recurrent Neural Network For Image Generationcited by 191
  6. 2018BAM: Bottleneck Attention Modulecited by 172
  7. 2021How to represent part-whole hierarchies in a neural networkcited by 123
  8. 2013Multi-digit Number Recognition from Street View Imagery using Deep Convolutional Neural Networkscited by 84
  9. 2021Differentiable Patch Selection for Image Recognitioncited by 68
  10. 2020An interpretable classifier for high-resolution breast cancer screening images utilizing weakly supervised localizationcited by 43
  11. 2019Saccader: Improving Accuracy of Hard Attention Models for Visioncited by 35
  12. 2016Feedback Networkscited by 29
  13. 2019Disentangling neural mechanisms for perceptual groupingcited by 28
  14. 2019Towards Robust Image Classification Using Sequential Attention Modelscited by 27
  15. 2020Capsules with Inverted Dot-Product Attention Routingcited by 25
  16. 2020Going in circles is the way forward: the role of recurrence in visual inferencecited by 19
  17. 2021Hard-Attention for Scalable Image Classificationcited by 18
  18. 2014Deep Networks with Internal Selective Attention through Feedback Connectionscited by 17
  19. 2017Learning to detect chest radiographs containing lung nodules using visual attention networkscited by 16
  20. 2021Capsule Network is Not More Robust than Convolutional Networkcited by 12
  21. 2014Attentional Neural Network: Feature Selection Using Cognitive Feedbackcited by 8
  22. 2017Hierarchical Attentive Recurrent Trackingcited by 7
  23. 2020Deflecting Adversarial Attackscited by 7
  24. 2019Capsule Networks with Max-Min Normalizationcited by 6
  25. 2015RATM: Recurrent Attentive Tracking Modelcited by 4
  26. 2020Discovering Dynamic Salient Regions for Spatio-Temporal Graph Neural Networkscited by 3
  27. 2017Learning with Rethinking: Recurrently Improving Convolutional Neural Networks through Feedbackcited by 2

Built on by

Cited by year

Citing papers per year: 2021: 1; 2022: 1.

2 papers.

  1. 2021Slot-VPS: Object-centric Representation Learning for Video Panoptic Segmentationcited by 24
  2. 2022Reconstruction-guided attention improves the robustness and shape processing of neural networkscited by 0