AI Paper Atlas

Research map / Detection and segmentation / CNN features

SIREN: Shaping Representations for Detecting Out-of-Distribution Objects

Xuefeng Du, Gabriel Gozum, Yifei Ming, Yixuan Li

NeurIPS 2022 · Detection and segmentation › CNN features · approach: Box regression and loss design · 10 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 SIREN hyperspherical embedding shaping, KNN scoreDeformable-DETR · PASCAL-VOC / MS-COCOAUROC %: 62.15 -> 78.23Improved
use vMF score on shaped embeddings instead of MahalanobisDeformable-DETR · PASCAL-VOC / MS-COCOAUROC %: 50.28 -> 76.10Improved
SIREN-KNN vs OW-DETRDeformable-DETR · PASCAL-VOC / MS-COCOAUROC %: 55.70 -> 78.23Improved
SIREN-KNN vs KNNDeformable-DETR · BDD100K / MS-COCOFPR95 %: 64.75 -> 53.97Improved
SIREN-KNN vs VOSFaster R-CNN · PASCAL-VOC / MS-COCOAUROC %: 88.70 -> 89.67Improved
use learnable kappa instead of fixed kappa=10 (ablation)SIREN-KNN · PASCAL-VOC / MS-COCOAUROC %: 77.67 -> 78.23Improved
use learned kappa instead of kappa estimated from embeddings (ablation)SIREN-vMF · PASCAL-VOC / MS-COCOAUROC %: 73.03 -> 76.10Improved
use nonlinear projection head instead of linear (ablation)SIREN-vMF · PASCAL-VOC / MS-COCOAUROC %: 71.73 -> 76.10Improved
use nonlinear projection head instead of linear (ablation)SIREN-KNN · PASCAL-VOC / MS-COCOAUROC %: 78.02 -> 78.23Improved
use KNN distance instead of vMF score (ablation)SIREN · PASCAL-VOC / MS-COCOAUROC %: 76.10 -> 78.23Improved
reduce hypersphere dimension to 8 (ablation)SIREN-KNN · PASCAL-VOC / MS-COCOAUROC %: 78.23 -> 77.74Hurt
increase hypersphere dimension to 80 (ablation)SIREN-KNN · PASCAL-VOC / MS-COCOAUROC %: 78.23 -> 76.76Hurt

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