Research map / Vision-language and multimodal / Human and embodied modeling
A Simple and Robust Correlation Filtering Method for Text-Based Person Search
Wei Suo, Mengyang Sun, Kai Niu, Yiqi Gao, Peng Wang, Yanning Zhang, Qi Wu
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 |
|---|---|---|---|
| add image and text denoising filters to the baseline (ablation) | Global+Local baseline · CUHK-PEDES | Top-1 accuracy: boost the baseline accuracy by 3.40% and 1.89% on Top-1 | Improved |
| add image and text dictionary filters to the baseline (ablation) | Global+Local baseline · CUHK-PEDES | Top-1 accuracy: outperform the baseline by 3.43% and 1.92% on Top-1 | Improved |
| combine denoising and dictionary filters (ablation) | Global+Local baseline · CUHK-PEDES | Top-1 accuracy: 59.76 -> 64.04 | Improved |
| use denoising filters instead of space and word attention (ablation) | SRCF-BERT · CUHK-PEDES | Top-1 accuracy: 61.89 -> 64.04 | Improved |
| use dictionary filters instead of contextual non-local attention (ablation) | SRCF-BERT · CUHK-PEDES | Top-1 accuracy: 60.85 -> 64.04 | Improved |
| update denoising filters by back propagation rather than moving average (ablation) | SRCF-BERT · CUHK-PEDES | Top-1 accuracy: 62.69 -> 64.04 | Improved |
| update dictionary filters by moving average rather than back propagation (ablation) | SRCF-BERT · CUHK-PEDES | Top-1 accuracy: 62.15 -> 64.04 | Improved |
| use separate filters per modality instead of shared denoising filters (ablation) | SRCF-BERT · CUHK-PEDES | Top-1 accuracy: 61.99 -> 64.04 | Improved |
| use separate filters per modality instead of shared dictionary filters (ablation) | SRCF-BERT · CUHK-PEDES | Top-1 accuracy: 61.45 -> 64.04 | Improved |
| use max pooling instead of avg-pooling (ablation) | SRCF-BERT · CUHK-PEDES | Top-1 accuracy: 63.97 -> 64.04 | No clear effect |
| add mutual-exclusion loss (ablation) | SRCF-BERT · CUHK-PEDES | Top-1 accuracy: 62.44 -> 64.04 | Improved |
| use SRCF instead of NAFS under random rotating test galleries | SRCF-BERT · CUHK-PEDES | Top-1 accuracy: 49.90 -> 59.23 | Improved |
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.
- ViTAA — ViTAA: Visual-Textual Attributes Alignment in Person Search by Natural Language (ECCV 2020) · better on 2 of 2 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
- Exploring the Potential of Large Vision-Language Models for Unsupervised Text-Based Person Retrieval (AAAI 2025 · similarity 0.88)
- Diverse Person: Customize Your Own Dataset for Text-Based Person Search (AAAI 2024 · similarity 0.88)
- Adaptive Uncertainty-Based Learning for Text-Based Person Retrieval (AAAI 2024 · similarity 0.87)
- Automatic Synthetic Data and Fine-grained Adaptive Feature Alignment for Composed Person Retrieval (NeurIPS 2025 · similarity 0.87)
- An Empirical Study of CLIP for Text-Based Person Search (AAAI 2024 · similarity 0.86)
- Causality-Inspired Invariant Representation Learning for Text-Based Person Retrieval (AAAI 2024 · similarity 0.86)
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.
