Research map / Vision-language and multimodal / Human and embodied modeling
Matching Handwritten Document Images
Praveen Krishnan, C. V. Jawahar
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 |
|---|---|---|---|
| replace KCSR features with HWNet CNN features | word spotting · GW | mAP: 0.9290 -> 0.9484 | Improved |
| replace KCSR features with HWNet CNN features | word spotting · IAM | mAP: 0.5573 -> 0.8061 | Improved |
| use stem-normalized features instead of plain CNN features (ablation) | word spotting · IAM inexact evaluation | mAP: 0.7170 -> 0.7443 | Improved |
| use stem-normalized features instead of plain CNN features (ablation) | word spotting · IAM exact evaluation | mAP: 0.8061 -> 0.7955 | Hurt |
| use MODS matching instead of SWM with CNN features (ablation) | document similarity · DocSim | nDCG@99: 0.8569 -> 0.8993 | Improved |
| use MODS matching instead of SWM with CNN features (ablation) | document similarity · DocSim | AUC: 0.9465 -> 0.9720 | Improved |
| use MODS matching instead of SWM with KCSR features (ablation) | document similarity · DocSim | nDCG@99: 0.7968 -> 0.8444 | Improved |
| use MODS matching instead of SWM with KCSR features (ablation) | document similarity · DocSim | AUC: 0.8231 -> 0.8302 | Improved |
| use CNN features instead of KCSR features in MODS (ablation) | document similarity · DocSim | nDCG@99: 0.8444 -> 0.8993 | Improved |
| use CNN features with MODS instead of SIFT bag of words | document similarity · DocSim | nDCG@99: 0.6128 -> 0.8993 | Improved |
Closest papers by meaning
- Text Matching as Image Recognition (AAAI 2016 · similarity 0.88)
- Expressing Objects Just Like Words: Recurrent Visual Embedding for Image-Text Matching (AAAI 2020 · similarity 0.86)
- Handwritten Text Recognition: A Survey (TPAMI 2025 · similarity 0.86)
- Neural Ctrl-F: Segmentation-Free Query-By-String Word Spotting in Handwritten Manuscript Collections (ICCV 2017 · similarity 0.85)
- An Unsupervised Sampling Approach for Image-Sentence Matching Using Document-level Structural Information (AAAI 2021 · similarity 0.84)
- Achieving Ensemble-Like Performance in a Single Model: A Feature Diversification Framework for Image-Text Matching (AAAI 2025 · similarity 0.82)
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.
