AI Paper Atlas

Research map / Detection and segmentation / CNN features

Effective Lymph Nodes Detection in CT Scans Using Location Debiased Query Selection and Contrastive Query Representation in Transformer

Qinji Yu, Yirui Wang, Ke Yan, Haoshen Li, Dazhou Guo, Li Zhang, Na Shen, Qifeng Wang, Xiaowei Ding, Le Lu, Xianghua Ye, Dakai Jin

ECCV 2024 · Detection and segmentation › CNN features · approach: Domain-specific segmentation and detection · 13 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-scale 2.5D feature fusion to Mask DINO (ablation)Mask DINO · LN detection internal testaverage recall (%): 51.32 -> 52.40Improved
add query contrastive learning on top of 2.5D fusion (ablation)LN-DETR · LN detection internal testaverage recall (%): 52.40 -> 53.44Improved
add location debiased query selection on top of 2.5D fusion (ablation)LN-DETR · LN detection internal testaverage recall (%): 52.40 -> 54.51Improved
combine debiased query selection and contrastive learning (ablation)LN-DETR · LN detection internal testaverage recall (%): 52.40 -> 56.27Improved
use LN-DETR instead of Mask DINO on external testMask DINO · LN detection external testaverage recall (%): 46.89 -> 52.04Improved
apply detection transformer (Mask DINO) directly without LN-specific changes versus CNN-based MULANMask DINO · LN detection external testaverage recall (%): 48.03 -> 46.89Hurt
detect by instance segmentation (Mask2Former/MP-Former) instead of direct detectionLN detection · LN detection internal testaverage recall (%): 45.11 and 47.69 vs 51.32 (Mask DINO)Hurt
use LN-DETR on enlarged LNs (>=10mm)Mask DINO · LN size >= 10mmaverage recall (%): 72.01 -> 77.15Improved
apply LN-DETR to universal lesion detection on DeepLesionLN-DETR · NIH DeepLesion testaverage recall (%): 88.46 vs P3D* 88.55Mixed

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.

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

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