Research map / Detection and segmentation / Open-set and few-shot detection
"Unitail: Detecting, Reading, and Matching in Retail Scene"
Fangyi Chen, Han Zhang, Zaiwang Li, Jiachen Dou, Shentong Mo, Hao Chen, Yongxin Zhang, Uzair Ahmed, Chenchen Zhu, Marios Savvides
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 Quad-Centerness (ablation) | RetailDet base · Unitail-Det val | mAP: 58.3 -> 60.4 | Improved |
| add Soft Scale assignment (ablation) | RetailDet + QC · Unitail-Det val | mAP: 60.4 -> 61.4 | Improved |
| add Corner Refinement Module (ablation) | RetailDet + QC + SS · Unitail-Det val | mAP: 61.4 -> 63.8 | Improved |
| add Corner Refinement Module (ablation) | RetailDet + QC + SS · Unitail-Det val | AP75: 72.6 -> 75.8 | Improved |
| train text detector on Unitail instead of SynthText+IC15 | DBNet · Unitail-OCR text detection | hmean: 0.666 -> 0.819 | Improved |
| train text detector on Unitail instead of SynthText+IC15 | FCENet · Unitail-OCR text detection | hmean: 0.538 -> 0.825 | Improved |
| add Unitail to text recognition training data | ABINet · Unitail-OCR text recognition | accuracy: 69.2 -> 77.2 | Improved |
| add Unitail to text recognition training data | CRNN · Unitail-OCR text recognition | accuracy: 40.0 -> 51.4 | Improved |
| add text matching to visual classifier | EfficientNetV2 · Unitail product matching (Hard) | accuracy: 56.49 -> 59.24 | Improved |
| add text matching to visual classifier | ResNet101 · Unitail product matching (Hard) | accuracy: 58.37 -> 60.52 | Improved |
| add text matching to visual classifier | ResNet101 · Unitail product matching (General) | accuracy: 85.03 -> 86.19 | Improved |
| use ResNet101 instead of ResNet50 backbone | RetailDet · Unitail-Det | g-mAP: 54.7 -> 57.1 | 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.
- PANet — PANet: Few-Shot Image Semantic Segmentation With Prototype Alignment (ICCV 2019) · better on 1 of 1 shared result
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
- ScaleNet: Guiding Object Proposal Generation in Supermarkets and Beyond (ICCV 2017 · similarity 0.83)
- MVTec D2S: Densely Segmented Supermarket Dataset (ECCV 2018 · similarity 0.83)
- Rethinking Object Detection in Retail Stores (AAAI 2021 · similarity 0.82)
- LLMs Meet VLMs: Boost Open Vocabulary Object Detection with Fine-grained Descriptors (ICLR 2024 · similarity 0.80)
- OmniGlue: Generalizable Feature Matching with Foundation Model Guidance (CVPR 2024 · similarity 0.80)
- MANTA: A Large-Scale Multi-View and Visual-Text Anomaly Detection Dataset for Tiny Objects (CVPR 2025 · similarity 0.79)
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.
