Prime Sample Attention in Object Detection

TL;DR

This paper introduces the concept of Prime Samples in object detection, proposing a new sampling strategy called PISA that emphasizes these key samples to improve detection performance over traditional methods.

Contribution

It identifies the importance of Prime Samples in training and develops PISA, a strategy that enhances object detection accuracy by focusing on these samples.

Findings

PISA outperforms random sampling and hard mining methods on MSCOCO.

Focusing on Prime Samples improves detection accuracy by around 2%.

PISA is effective for both single-stage and two-stage detectors.

Abstract

It is a common paradigm in object detection frameworks to treat all samples equally and target at maximizing the performance on average. In this work, we revisit this paradigm through a careful study on how different samples contribute to the overall performance measured in terms of mAP. Our study suggests that the samples in each mini-batch are neither independent nor equally important, and therefore a better classifier on average does not necessarily mean higher mAP. Motivated by this study, we propose the notion of Prime Samples, those that play a key role in driving the detection performance. We further develop a simple yet effective sampling and learning strategy called PrIme Sample Attention (PISA) that directs the focus of the training process towards such samples. Our experiments demonstrate that it is often more effective to focus on prime samples than hard samples when training a detector. Particularly, On the MSCOCO dataset, PISA outperforms the random sampling baseline and hard mining schemes, e.g., OHEM and Focal Loss, consistently by around 2% on both single-stage and two-stage detectors, even with a strong backbone ResNeXt-101.