Research map / Detection and segmentation / CNN features
TableSense: Spreadsheet Table Detection with Convolutional Neural Networks
Haoyu Dong, Shijie Liu, Shi Han, Zhouyu Fu, Dongmei Zhang
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 Precise Bounding box Regression (PBR) module on top of BBR (ablation) | TableSense · WebSheet400 | EoB-2 recall: 78.4 -> 91.3 | Improved |
| add PBR module on top of BBR (ablation) | TableSense · WebSheet400 | EoB-2 precision: 68.1 -> 86.5 | Improved |
| use full 20-feature cell featurization instead of binary valued features only (ablation) | TableSense · WebSheet400 | EoB-2 recall: 69.4 -> 91.3 | Improved |
| use full feature set instead of value string features only (ablation) | TableSense · WebSheet400 | EoB-2 precision: 77.3 -> 86.5 | Improved |
| use full feature set instead of value string features only (ablation) | TableSense · WebSheet400 | EoB-0 recall: 69.5 -> 80.8 | Improved |
| use TableSense (customized CNN with PBR) instead of Mask R-CNN | spreadsheet table detection · WebSheet400 | EoB-0 precision: 40.2 -> 78.1 | Improved |
| use TableSense instead of Mask R-CNN | spreadsheet table detection · WebSheet400 | EoB-2 recall: 75.5 -> 91.3 | Improved |
| use TableSense instead of region-growth | spreadsheet table detection · WebSheet400 | EoB-2 precision: 52.2 -> 86.5 | Improved |
| use two-stage Mask R-CNN instead of one-stage YOLO-v3 | baseline detectors · WebSheet400 | EoB-2 recall: 63.4 -> 75.5 | Improved |
| use model from 7th active learning iteration instead of 3rd for uncertainty-based sheet selection | uncertainty metric · WebSheet400 | selection accuracy on sheets with errors: 81.9% -> 86.4% | 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.
- Faster R-CNN — Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks (TPAMI 2016) · 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
- TGRNet: A Table Graph Reconstruction Network for Table Structure Recognition (ICCV 2021 · similarity 0.86)
- TDeLTA: A Light-Weight and Robust Table Detection Method Based on Learning Text Arrangement (AAAI 2024 · similarity 0.86)
- TABLET: A Large-Scale Dataset for Robust Visual Table Understanding (ICLR 2026 · similarity 0.82)
- TRivia: Self-supervised Fine-tuning of Vision-Language Models for Table Recognition (CVPR 2026 · similarity 0.82)
- Grab What You Need: Rethinking Complex Table Structure Recognition with Flexible Components Deliberation (AAAI 2024 · similarity 0.82)
- MMTU: A Massive Multi-Task Table Understanding and Reasoning Benchmark (NeurIPS 2025 · similarity 0.81)
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.
