Recovering hard-to-find object instances by sampling context-based object proposals

TL;DR

This paper introduces a post-detection sampling method that explores images to recover missed object detections, leveraging spatial relations and higher-order group relations to improve recall efficiently.

Contribution

It presents novel strategies for sampling object proposals based on spatial and group relations, enhancing detection recall after initial detection.

Findings

Improved recall on KITTI dataset using relation-based proposal strategies

Achieved higher detection recall with fewer object proposals

Demonstrated effectiveness of higher-order relation discovery in object proposal sampling

Abstract

In this paper we focus on improving object detection performance in terms of recall. We propose a post-detection stage during which we explore the image with the objective of recovering missed detections. This exploration is performed by sampling object proposals in the image. We analyze four different strategies to perform this sampling, giving special attention to strategies that exploit spatial relations between objects. In addition, we propose a novel method to discover higher-order relations between groups of objects. Experiments on the challenging KITTI dataset show that our proposed relations-based proposal generation strategies can help improving recall at the cost of a relatively low amount of object proposals.