Research map / Detection and segmentation / Weakly supervised detection and segmentation
Weakly Supervised Coupled Networks for Visual Sentiment Analysis
Jufeng Yang, Dongyu She, Yu-Kun Lai, Paul L. Rosin, Ming-Hsuan Yang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use WSCNet instead of WILDCAT | ResNet-101 · FI test | accuracy: 70.07 vs 67.03 | Improved |
| Use WSCNet instead of WILDCAT | ResNet-101 · Flickr test | accuracy: 81.36 vs 80.67 | Improved |
| Use WSCNet instead of WILDCAT | ResNet-101 · Instagram test | accuracy: 81.81 vs 80.31 | Improved |
| Use WSCNet instead of WILDCAT | ResNet-101 · EmotionROI test | accuracy: 58.25 vs 55.05 | Improved |
| Use WSCNet instead of SPN | ResNet-101 · Twitter I test | accuracy: 84.25 vs 81.67 | Improved |
| Use WSCNet instead of Sun et al. | ResNet-101 · Twitter II test | accuracy: 81.35 vs 80.84 | Improved |
| Use 4 feature maps per sentiment class rather than 1 (ablation) | WSCNet · FI test | accuracy: 68.23 -> 70.07 | Improved |
| Use more than 4 feature maps per sentiment class (ablation) | WSCNet · FI test | accuracy: 70.07 -> 68.80; 70.07 -> 67.19 | Hurt |
| Vary the number of feature maps from 1 to 16 (ablation) | WSCNet · Flickr test | accuracy: 81.46; 81.87; 81.36; 81.15; 81.98 | Mixed |
| Vary the number of feature maps from 1 to 16 (ablation) | WSCNet · Instagram test | accuracy: 79.67; 79.24; 81.80; 79.60; 78.53 | Mixed |
| Add multiple feature maps (k=4) and sentiment-map coupling instead of the baseline without coupling (ablation) | WSCNet · FI test | accuracy: 66.57 -> 70.07 | Improved |
| Add sentiment-map guidance to the k=4 model (ablation) | WSCNet · FI test | accuracy: 67.96 -> 67.69 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- WILDCAT — WILDCAT: Weakly Supervised Learning of Deep ConvNets for Image Classification, Pointwise Localization and Segmentation (CVPR 2017) · better on 4 of 4 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Probing Sentiment-Oriented Pre-Training Inspired by Human Sentiment Perception Mechanism (CVPR 2023 · similarity 0.89)
- Visual Sentiment Analysis by Attending on Local Image Regions (AAAI 2017 · similarity 0.88)
- VistaNet: Visual Aspect Attention Network for Multimodal Sentiment Analysis (AAAI 2019 · similarity 0.86)
- Learning Visual Sentiment Distributions via Augmented Conditional Probability Neural Network (AAAI 2017 · similarity 0.86)
- Kill Two Birds With One Stone: Weakly-Supervised Neural Network for Image Annotation and Tag Refinement (AAAI 2018 · similarity 0.86)
- TextGT: A Double-View Graph Transformer on Text for Aspect-Based Sentiment Analysis (AAAI 2024 · similarity 0.83)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
