Research map / Image generation / Image restoration
Fractal Dimension Invariant Filtering and Its CNN-Based Implementation
Hongteng Xu, Junchi Yan, Nils Persson, Weiyao Lin, Hongyuan Zha
Check 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 |
|---|---|---|---|
| apply unsupervised FDIF filtering with thresholding instead of traditional non-learning detectors | curve detection · AFM image benchmark | ODS: 0.058 -> 0.717; 0.629 -> 0.717 | Improved |
| use supervised FraCNN features with logistic regression rather than HED | curve detection · AFM image benchmark | ODS: 0.722 -> 0.743 | Improved |
| use supervised FraCNN features with logistic regression rather than HED | curve detection · AFM image benchmark | AP: 0.784 -> 0.730 | Hurt |
| increase FDIF iterations from 3 to 6 (ablation) | FDIF + BP · AFM image benchmark | ODS: 0.717 -> 0.715 | Hurt |
| increase FDIF iterations from 3 to 6 (ablation) | FDIF + LR · AFM image benchmark | OIS: 0.770 -> 0.767 | Hurt |
| increase FraCNN depth from 6 to 12 (ablation) | FraCNN + BP · AFM image benchmark | ODS: 0.691 -> 0.689 | Hurt |
| increase FraCNN depth from 6 to 12 (ablation) | FraCNN + LR · AFM image benchmark | AP: 0.730 -> 0.718 | Hurt |
| replace adaptive direction selection with a predefined anisotropic filter bank | FDIF curve detector · AFM image benchmark | AP: 0.699 -> 0.708 | Mixed |
| use FDIF rather than HED on Brodatz texture curves | curve detection · Brodatz texture images | ODS: 0.518 -> 0.530 | Improved |
| use FDIF rather than HED on Brodatz texture curves | curve detection · Brodatz texture images | AP: 0.791 -> 0.803 | Improved |
Closest papers by meaning
- Encoding Spatial Distribution of Convolutional Features for Texture Representation (NeurIPS 2021 · similarity 0.87)
- Fractal-Domain Vision Graph Neural Network for Remote Sensing Ground Target Classification (TPAMI 2026 · similarity 0.86)
- MFSR: Multi-Fractal Feature for Super-Resolution Reconstruction With Fine Details Recovery (TPAMI 2026 · similarity 0.86)
- Learning Fractals by Gradient Descent (AAAI 2023 · similarity 0.84)
- FGM-HD: Boosting Generation Diversity of Fractal Generative Models through Hausdorff Dimension Induction (AAAI 2026 · similarity 0.84)
- Fractal Calibration for Long-tailed Object Detection (CVPR 2025 · 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.
