Learning Instance-Aware Object Detection Using Determinantal Point Processes

TL;DR

This paper introduces IDNet, a novel instance-aware detection network that leverages determinantal point processes to improve detection of overlapped objects, outperforming existing methods on standard datasets.

Contribution

The paper proposes a new network architecture, IDNet, that learns to distinguish objects using pairwise similarities and applies determinantal point process inference for better candidate selection.

Findings

Significant improvement in overlapped object detection accuracy.

Outperforms state-of-the-art methods on PASCAL VOC and MS COCO.

Effective handling of overlapping objects in detection tasks.

Abstract

Recent object detectors find instances while categorizing candidate regions. As each region is evaluated independently, the number of candidate regions from a detector is usually larger than the number of objects. Since the final goal of detection is to assign a single detection to each object, a heuristic algorithm, such as non-maximum suppression (NMS), is used to select a single bounding box for an object. While simple heuristic algorithms are effective for stand-alone objects, they can fail to detect overlapped objects. In this paper, we address this issue by training a network to distinguish different objects using the relationship between candidate boxes. We propose an instance-aware detection network (IDNet), which can learn to extract features from candidate regions and measure their similarities. Based on pairwise similarities and detection qualities, the IDNet selects a subset of candidate bounding boxes using instance-aware determinantal point process inference (IDPP). Extensive experiments demonstrate that the proposed algorithm achieves significant improvements for detecting overlapped objects compared to existing state-of-the-art detection methods on the PASCAL VOC and MS COCO datasets.