AI Paper Atlas

Research map / Detection and segmentation / Open-set and few-shot detection

MosaicOS: A Simple and Effective Use of Object-Centric Images for Long-Tailed Object Detection

Cheng Zhang, Tai-Yu Pan, Yandong Li, Hexiang Hu, Dong Xuan, Soravit Changpinyo, Boqing Gong, Wei-Lun Chao

ICCV 2021 · Detection and segmentation › Open-set and few-shot detection · approach: Open-set and unknown object detection · 42 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 2x2 mosaicking of object-centric images (ablation)MosaicOS (two-stage, F) · LVIS v0.5 valAP^b_r: 16.97 -> 18.76Improved
stitch images from different classes instead of the same class (ablation)MosaicOS · LVIS v0.5 valAP^b_r: 18.76 -> 19.73Improved
use two-stage fine-tuning instead of single-stage joint fine-tuning (ablation)MosaicOS · LVIS v0.5 valAP^b: 20.09 -> 24.75Improved
replace detector-predicted labels with the image class label on pseudo boxes (ablation)MosaicOS (pre-trained detector pseudo-labels) · LVIS v0.5 valAP^b_r: 13.93 -> 17.45Improved
add image class labels to calibrated-detector pseudo-boxes (ablation)MosaicOS (calibrated detector) · LVIS v0.5 valAP^b_r: 13.13 -> 19.31Improved
add 100 Google images per class to ImageNet object-centric imagesMosaicOS · LVIS v0.5 valAP^b_r: 19.73 -> 20.25Improved
optimize RPN and box regression losses in addition to classification loss in stage one (ablation)MosaicOS · LVIS v0.5 valAP^b_r: 18.87 -> 19.73Improved
use LORE pseudo-boxes instead of six fixed locations (two-stage) (ablation)MosaicOS · LVIS v0.5 valAP^b_r: 19.73 -> 20.06No clear effect
use 3x3 instead of 2x2 mosaic (ablation)MosaicOS (same-class mosaic) · LVIS v0.5 valAP^b_r: 18.76 -> 18.14No clear effect
stitch object-centric images instead of stitching scene-centric LVIS images (ablation)Faster R-CNN* · LVIS v0.5 valAP^b_r: 13.99 -> 19.73Improved
use ImageNet object-centric images instead of crawled Flickr imagesMosaicOS · LVIS v0.5 valAP^b_r: 16.17 -> 19.73Improved
add MosaicOS fine-tuning on classes already seen in ImageNet-1K backbone pre-trainingFaster R-CNN* · LVIS v0.5 val (176 overlapping classes)AP^b_r: 14.78 -> 21.16Improved

Later papers that compared with this method

2 later accepted papers report MosaicOS as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.

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