Research map / Detection and segmentation / CNN features
Rethinking the Defocus Blur Detection Problem and A Real-Time Deep DBD Model
Ning Zhang, Junchi Yan
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 hard mining to BCE+self-supervision (ablation) | BCE+self-supervision U-Net · CUHK | F-measure: 0.933 -> 0.944 | Improved |
| add hard mining to BCE+self-supervision (ablation) | BCE+self-supervision U-Net · CUHK | MAE: 0.060 -> 0.053 | Improved |
| add hard mining to BCE+self-supervision (ablation) | BCE+self-supervision U-Net · DUT | F-measure: 0.811 -> 0.828 | Improved |
| add hard mining to BCE+self-supervision (ablation) | BCE+self-supervision U-Net · DUT | MAE: 0.127 -> 0.115 | Improved |
| add self-supervision loss to BCE (ablation) | BCE U-Net · CUHK | F-measure: 0.927 -> 0.933 | Improved |
| add self-supervision loss to BCE (ablation) | BCE U-Net · CUHK | MAE: 0.066 -> 0.060 | Improved |
| add self-supervision loss to BCE (ablation) | BCE U-Net · DUT | F-measure: 0.793 -> 0.811 | Improved |
| add self-supervision loss to BCE (ablation) | BCE U-Net · DUT | MAE: 0.140 -> 0.127 | Improved |
| add image gradient input to U-Net (ablation) | BCE U-Net · CUHK | F-measure: 0.918 -> 0.927 | Improved |
| add image gradient input to U-Net (ablation) | BCE U-Net · DUT | F-measure: 0.754 -> 0.793 | Improved |
| add image gradient input to U-Net (ablation) | BCE U-Net · FPS | speed: 112.35 -> 58.82 | Hurt |
| vary self-supervision weight lambda_s from 0.0 to 0.5 (ablation) | BCE+self-supervision · CUHK / DUT | F-measure and MAE: CUHK F 0.927 -> 0.933 at 0.2; DUT F 0.793 -> 0.811 at 0.2 | No clear effect |
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.
- DeFusionNet — DeFusionNET: Defocus Blur Detection via Recurrently Fusing and Refining Multi-Scale Deep Features (CVPR 2019) · better on 5 of 5 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
- Defocus Blur Detection via Multi-Stream Bottom-Top-Bottom Fully Convolutional Network (CVPR 2018 · similarity 0.93)
- DeFusionNET: Defocus Blur Detection via Recurrently Fusing and Refining Multi-Scale Deep Features (CVPR 2019 · similarity 0.92)
- DeFusionNET: Defocus Blur Detection via Recurrently Fusing and Refining Discriminative Multi-Scale Deep Features (TPAMI 2020 · similarity 0.92)
- Defocus Blur Detection via Depth Distillation (ECCV 2020 · similarity 0.92)
- Defocus Blur Detection via Multi-Stream Bottom-Top-Bottom Network (TPAMI 2019 · similarity 0.91)
- Multi-Focus Image Fusion via Explicit Defocus Blur Modelling (AAAI 2025 · similarity 0.85)
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.
