atlas

Research map / Detection and segmentation / CNN features

FastMask: Segment Multi-Scale Object Candidates in One Shot

CVPR 2017 · Detection and segmentation › CNN features · approach: Region-based and query-based detectors · 31 citations (OpenAlex)

Check your idea against 108,967 papers

AI-extracted summary. Written by AI from the paper's title only (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 a residual component to the average-pooling neck (ablation)FastMask · COCOAR@100: 27.9 -> 29.3Helped
use max pooling instead of the residual neck (ablation)FastMask · COCOAR@100: 29.3 -> 27.8Hurt
use a feed-forward neck instead of the residual neck (ablation)FastMask · COCOAR@100: 29.3 -> 27.1Hurt
add spatial attention to the decoding head (ablation)FastMask · COCOAR@100: 24.8 -> 29.3Helped
add spatial attention to the decoding head (ablation)FastMask · COCOAR@10: 12.7 -> 15.2Helped
reduce the feature pyramid from 8 scales to 4 (ablation)FastMask · COCOAR@100: 31.3 -> 26.6Hurt
reduce the feature pyramid from 8 scales to 4 (ablation)FastMask · COCOSpeed: 0.26s -> 0.14sHelped
use a two-stream feature pyramid rather than image-pyramid inferenceFastMask · COCO segmentation proposalsSpeed: 0.46s -> 0.26sHelped

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