Research map / Architectures and efficiency / CNN design
Quadtree Convolutional Neural Networks
Pradeep Kumar Jayaraman, Jianhan Mei, Jianfei Cai, Jianmin Zheng
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 dense CNN with quadtree CNN | CNN · MNIST | #MACC x10^6: 18.36 -> 6.1 | Improved |
| replace dense CNN with quadtree CNN | CNN · EMNIST Balanced | #MACC x10^6: 18.36 -> 9.03 | Improved |
| replace dense CNN with quadtree CNN | CNN · CASIA-HWDB1.1 | #MACC x10^6: 229.34 -> 136.7 | Improved |
| replace dense CNN with quadtree CNN | CNN · TU-Berlin | #MACC x10^6: 837.53 -> 257.07 | Improved |
| replace dense CNN with quadtree CNN | CNN · MNIST | accuracy (%): 99.48 -> 99.39 | Hurt |
| replace dense CNN with quadtree CNN | CNN · EMNIST Balanced | accuracy (%): 89.53 -> 88.89 | Hurt |
| replace dense CNN with quadtree CNN | CNN · CASIA-HWDB1.1 | accuracy (%): 97.44 -> 97.36 | No clear effect |
| replace dense CNN with quadtree CNN | CNN · TU-Berlin | accuracy (%): 61.13 -> 62.44 | No clear effect |
| replace dense CNN with quadtree CNN | CNN · Sketch simplification 256x256 | memory (MB): 127.14 -> 13.5 | Improved |
| replace dense CNN with quadtree CNN | CNN · Sketch simplification 1024x1024 | memory (MB): 2034.24 -> 97.9 | Improved |
| replace dense CNN with quadtree CNN | CNN · Sketch simplification 1024x1024 | #MACC x10^8: 11596.21 -> 917.14 | Improved |
| replace dense CNN with quadtree CNN in a deeper network | CNN · TU-Berlin | training time: one-third the training time | Improved |
Closest papers by meaning
- Quaternion Convolutional Neural Networks (ECCV 2018 · similarity 0.84)
- Quadtree Attention for Vision Transformers (ICLR 2022 · similarity 0.83)
- Deep Roots: Improving CNN Efficiency With Hierarchical Filter Groups (CVPR 2017 · similarity 0.83)
- SketchNet: Sketch Classification With Web Images (CVPR 2016 · similarity 0.83)
- Generalizing CNNs to graphs with learnable neighborhood quantization (NeurIPS 2024 · similarity 0.83)
- ViTree: Single-Path Neural Tree for Step-Wise Interpretable Fine-Grained Visual Categorization (AAAI 2024 · 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.
