AI Paper Atlas

Research map / Detection and segmentation / CNN features

A Cephalometric Landmark Regression Method based on Dual-encoder for High-resolution X-ray Image

Chao Dai, yang wang, Chaolin Huang, zhou jiakai, Qilin Xu, Minpeng Xu

ECCV 2024 · Detection and segmentation › CNN features · approach: Box regression and loss design · 5 citations (OpenAlex)

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 changedWhereMeasuredEffect
add multi-level features and heatmap (feature extractor module) (ablation)baseline regression · ISBI 2015 test1MRE (mm): 2.8974 -> 2.0125Improved
add reference encoder with RLE loss (ablation)baseline regression · ISBI 2015 test1MRE (mm): 2.8974 -> 1.6745Improved
add finetune encoder on top of reference encoder and feature extractor (ablation)Du-CeLR · ISBI 2015 test1MRE (mm): 1.1514 -> 1.0088Improved
feed more levels of fused feature maps (1 to 4 levels) to finetune encoder (ablation)Du-CeLR · ISBI 2015 test12mm SDR (%): 89.20 -> 89.51Improved
increase encoder dimension from 128 to 512 (ablation)Du-CeLR · ISBI 2015 test12mm SDR (%): 89.03 -> 89.51Improved
increase encoder dimension from 512 to 768 (ablation)Du-CeLR · ISBI 2015 test12mm SDR (%): 89.51 -> 89.47No clear effect
increase encoder layers from 1 to 4 (ablation)Du-CeLR · ISBI 2015 test12mm SDR (%): 87.23 -> 89.51Improved
increase encoder layers from 4 to 5 (ablation)Du-CeLR · ISBI 2015 test12mm SDR (%): 89.51 -> 89.48No clear effect
increase input resolution from 256 to 1024 (ablation)Du-CeLR · ISBI 2015 test12mm SDR (%): 84.56 -> 89.51Improved
increase input resolution from 1024 to 1280 (ablation)Du-CeLR · ISBI 2015 test12mm SDR (%): 89.51 -> 89.31Hurt
use 512x512 input instead of high-resolution regression cascadesDu-CeLR · ISBI 2015 test1GFLOPs: 220.2 -> 23.0Improved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy