ExpNet: A unified network for Expert-Level Classification
Not on the map. Why?
Abstract (arXiv)
Different from the general visual classification, some classification tasks are more challenging as they need the professional categories of the images. In the paper, we call them expert-level classification. Previous fine-grained vision classification (FGVC) has made many efforts on some of its specific sub-tasks. However, they are difficult to expand to the general cases which rely on the comprehensive analysis of part-global correlation and the hierarchical features interaction. In this paper, we propose Expert Network (ExpNet) to address the unique challenges of expert-level classification through a unified network. In ExpNet, we hierarchically decouple the part and context features and individually process them using a novel attentive mechanism, called Gaze-Shift. In each stage, Gaze-Shift produces a focal-part feature for the subsequent abstraction and memorizes a context-related embedding. Then we fuse the final focal embedding with all memorized context-related embedding to make the prediction. Such an architecture realizes the dual-track processing of partial and global information and hierarchical feature interactions. We conduct the experiments over three representative expert-level classification tasks: FGVC, disease classification, and artwork attributes classification. In these experiments, superior performance of our ExpNet is observed comparing to the state-of-the-arts in a wide range of fields, indicating the effectiveness and generalization of our ExpNet. The code will be made publicly available.
Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Builds on 48 of its 76 references that are in the map. First posted on arXiv 29 Nov 2022.
Builds on
48 of 76 references.
References · 48
- 2015Deep Residual Learning for Image Recognitioncited by 52,096
- 2014Adam: A Method for Stochastic Optimizationcited by 37,240
- 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 30,186
- 2020Training data-efficient image transformers & distillation through attentioncited by 4,586
- 2018CBAM: Convolutional Block Attention Modulecited by 2,477
- 2020Implicit Neural Representations with Periodic Activation Functionscited by 2,238
- 2013Fine-Grained Visual Classification of Aircraftcited by 1,726
- 2017Deformable Convolutional Networkscited by 1,488
- 2021CvT: Introducing Convolutions to Vision Transformerscited by 947
- 2021CrossViT: Cross-Attention Multi-Scale Vision Transformer for Image Classificationcited by 806
- 2013Network In Networkcited by 605
- 2021Neural Fields in Visual Computing and Beyondcited by 540
36 more
- 2021Conditional Positional Encodings for Vision Transformerscited by 406
- 2021FNet: Mixing Tokens with Fourier Transformscited by 347
- 2020COVID-Net: A Tailored Deep Convolutional Neural Network Design for Detection of COVID-19 Cases from Chest X-Ray Imagescited by 264
- 2021Global Filter Networks for Image Classificationcited by 260
- 2019Weakly Supervised Learning of Instance Segmentation with Inter-pixel Relationscited by 247
- 2016SCA-CNN: Spatial and Channel-wise Attention in Convolutional Networks for Image Captioningcited by 219
- 2015Large-scale Classification of Fine-Art Paintings: Learning The Right Metric on The Right Featurecited by 182
- 2021TransFG: A Transformer Architecture for Fine-grained Recognitioncited by 132
- 2021TS-CAM: Token Semantic Coupled Attention Map for Weakly Supervised Object Localizationcited by 123
- 2020Evaluating Weakly Supervised Object Localization Methods Rightcited by 105
- 2020Fine-Grained Visual Classification via Progressive Multi-Granularity Training of Jigsaw Patchescited by 92
- 2020Learning Attentive Pairwise Interaction for Fine-Grained Classificationcited by 84
- 2022Dual Cross-Attention Learning for Fine-Grained Visual Categorization and Object Re-Identificationcited by 78
- 2019Weakly Supervised Complementary Parts Models for Fine-Grained Image Classification from the Bottom Upcited by 66
- 2022CLIP-Art: Contrastive Pre-training for Fine-Grained Art Classificationcited by 65
- 2019Cross-X Learning for Fine-Grained Visual Categorizationcited by 61
- 2021An Image Patch is a Wave: Phase-Aware Vision MLPcited by 61
- 2022L2G: A Simple Local-to-Global Knowledge Transfer Framework for Weakly Supervised Semantic Segmentationcited by 57
- 2020Inter-Image Communication for Weakly Supervised Localizationcited by 52
- 2019Attention Convolutional Binary Neural Tree for Fine-Grained Visual Categorizationcited by 51
- 2018Disc-aware Ensemble Network for Glaucoma Screening from Fundus Imagecited by 39
- 2021Puzzle-CAM: Improved localization via matching partial and full featurescited by 37
- 2017Comparing deep neural networks against humans: object recognition when the signal gets weakercited by 34
- 2021RAMS-Trans: Recurrent Attention Multi-scale Transformer forFine-grained Image Recognitioncited by 34
- 2022Bridging the Gap between Classification and Localization for Weakly Supervised Object Localizationcited by 25
- 2021LCTR: On Awakening the Local Continuity of Transformer for Weakly Supervised Object Localizationcited by 24
- 2022Weakly Supervised Object Localization via Transformer with Implicit Spatial Calibrationcited by 20
- 2020Attribute Mix: Semantic Data Augmentation for Fine Grained Recognitioncited by 15
- 2022Dynamic MLP for Fine-Grained Image Classification by Leveraging Geographical and Temporal Informationcited by 15
- 2022Dynamic Group Transformer: A General Vision Transformer Backbone with Dynamic Group Attentioncited by 14
- 2020Leveraging Undiagnosed Data for Glaucoma Classification with Teacher-Student Learningcited by 12
- 2020Two-stage framework for optic disc localization and glaucoma classification in retinal fundus images using deep learningcited by 11
- 2021GCNBoost: Artwork Classification by Label Propagation through a Knowledge Graphcited by 11
- 2019The iMet Collection 2019 Challenge Datasetcited by 10
- 2022SeATrans: Learning Segmentation-Assisted diagnosis model via Transformercited by 8
- 2022RecurSeed and EdgePredictMix: Pseudo-Label Refinement Learning for Weakly Supervised Semantic Segmentation across Single- and Multi-Stage Frameworkscited by 1
Built on by
Citing papers per year: 2025: 1.
1 paper.