Unsupervised Detection and Segmentation of Identical Objects

Authors

Abstract

We address an unsupervised object detection and segmentation problem that goes beyond the conventional assumptions of one-to-one object correspondences or model-test settings between images. Our method can detect and segment identical objects directly from a single image or a handful of images without any supervision. To detect and segment all the object-level correspondences from the given images, a novel multi-layer match-growing method is proposed that starts from initial local feature matches and explores the images by intra-layer expansion and inter-layer merge. It estimates geometric relations between object entities and establishes `object correspondence networks' that connect matching objects. Experiments demonstrate robust performance of our method on challenging datasets.


paper thumbnail

Paper

CVPR 2010 paper. (pdf, 17.6MB)

Presentation

pdf file (slides), 10.4MB
pptx file (slides), 16.8MB

Dataset

dataset (zip), 26MB

Citation

Minsu Cho, Young Min Shin, and Kyoung Mu Lee, "Unsupervised Detection and Segmentation of Identical Objects," Proc. Computer Vision and Pattern Recognition (CVPR), 2010.



Experimental Results

Input
Object Networks
Local Matches
Input
Object Networks
Local Matches
Input
Object Networks
Local Matches
A Penta-tiled Image
Object Networks


Under Construction...