Research map / Detection and segmentation / CNN features
Synthetic Data for Text Localisation in Natural Images
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 multi-scale detection instead of single-scale detection (ablation) | FCRN · IC13 | maximum recall: more than 12% | Helped |
| Add high-quality multi-filter post-processing to multi-scale detections (ablation) | FCRN · IC13 | F-measure: +6% | Helped |
| Train Jaderberg's method on SynthText instead of its original training data (ablation) | Jaderberg text detector · IC13 | average precision: +0.8% in AP, +2.1% in maximum recall | Helped |
| Constrain synthetic text to local colour and texture regions rather than placing it randomly (ablation) | Synthetic text training data for FCRNall + multi-filt · SVT | maximum recall: +6.8% | Helped |
| Constrain synthetic text to local colour and texture regions rather than placing it randomly (ablation) | Synthetic text training data for FCRNall + multi-filt · SVT | average precision: +3.85% | Helped |
| Constrain synthetic text to local colour and texture regions rather than placing it randomly (ablation) | Synthetic text training data for FCRNall + multi-filt · SVT | maximum F-measure: +2.1% | Helped |
| Add perspective distortion matching local scene depth to locally constrained text placement (ablation) | Synthetic text training data for FCRNall + multi-filt · SVT | average precision: +0.75% | Helped |
| Add perspective distortion matching local scene depth to locally constrained text placement (ablation) | Synthetic text training data for FCRNall + multi-filt · SVT | maximum F-measure: +0.55% | Helped |
| Add perspective distortion matching local scene depth to locally constrained text placement (ablation) | Synthetic text training data for FCRNall + multi-filt · SVT | maximum recall: no change | No clear effect |
Closest papers by meaning
- Multi-Oriented Text Detection With Fully Convolutional Networks (CVPR 2016 · similarity 0.88)
- EAST: An Efficient and Accurate Scene Text Detector (CVPR 2017 · similarity 0.88)
- TextNeRF: A Novel Scene-Text Image Synthesis Method based on Neural Radiance Fields (CVPR 2024 · similarity 0.85)
- Scaling Laws of Synthetic Images for Model Training ... for Now (CVPR 2024 · similarity 0.84)
- LRANet: Towards Accurate and Efficient Scene Text Detection with Low-Rank Approximation Network (AAAI 2024 · similarity 0.84)
- Kernel Adaptive Convolution for Scene Text Detection via Distance Map Prediction (CVPR 2024 · similarity 0.84)
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.