Unsupervised Object Discovery and Localization in the Wild: Part-based Matching with Bottom-up Region Proposals

Abstract

This paper addresses unsupervised discovery and localization of dominant objects from a noisy image collection with multiple object classes. The setting of this problem is fully unsupervised, without even image-level annotations or any assumption of a single dominant class. This is far more general than typical colocalization, cosegmentation, or weakly-supervised localization tasks. We tackle the discovery and localization problem using a part-based region matching approach: We use off-the-shelf region proposals to form a set of candidate bounding boxes for objects and object parts. These regions are efficiently matched across images using a probabilistic Hough transform that evaluates the confidence for each candidate correspondence considering both appearance and spatial consistency. Dominant objects are discovered and localized by comparing the scores of candidate regions and selecting those that stand out over other regions containing them. Extensive experimental evaluations on standard benchmarks demonstrate that the proposed approach significantly outperforms the current state of the art in colocalization, and achieves robust object discovery in challenging mixed-class datasets.

CVPR 2015

References in corpus (1)

• Rich feature hierarchies for accurate object detection and semantic segmentation

Cited by in corpus (22)

• Identifying and Categorizing Anomalies in Retinal Imaging Data
• Learning Dense Correspondence via 3D-guided Cycle Consistency
• Relaxed Multiple-Instance SVM with Application to Object Discovery
• Unsupervised Object Discovery and Tracking in Video Collections
• Deep Descriptor Transforming for Image Co-Localization
• Ego-Object Discovery
• Re-ranking Object Proposals for Object Detection in Automatic Driving
• Mining Object Parts from CNNs via Active Question-Answering
• Deep Feature Factorization For Concept Discovery
• Group-wise Deep Co-saliency Detection
• Fusing Saliency Maps with Region Proposals for Unsupervised Object Localization
• Mining Interpretable AOG Representations from Convolutional Networks via Active Question Answering
• An evaluation of large-scale methods for image instance and class discovery
• cvpaper.challenge in 2015 - A review of CVPR2015 and DeepSurvey
• Spot On: Action Localization from Pointly-Supervised Proposals
• A regularization-based approach for unsupervised image segmentation
• A Comparative Study on Effects of Original and Pseudo Labels for Weakly Supervised Learning for Car Localization Problem
• cvpaper.challenge in 2016: Futuristic Computer Vision through 1,600 Papers Survey
• Generative Adversarial Networks for Unsupervised Object Co-localization
• Video Region Annotation with Sparse Bounding Boxes
• Unsupervised Action Proposal Ranking through Proposal Recombination
• Learning to Segment Object Candidates via Recursive Neural Networks