Prime Sample Attention in Object Detection

By Yuhang Cao, Kai Chen, Chen Change Loy, Dahua Lin. 2019

228 citations recorded by the source corpus.

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 result in 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, \eg~OHEM and Focal Loss, consistently by around 2\% on both single-stage and two-stage detectors, even with a strong backbone ResNeXt-101. Code is available at: \url{https://github.com/open-mmlab/mmdetection}.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

Citations in the corpus

Generalized Focal Loss: Learning Qualified and Distributed Bounding Boxes for Dense Object Detection (2020, background): “Instead of introducing an additional branch, PISA [2] and IoU-balance [28] assign different weights in the classification loss based on their localization qualities, aiming at enhancing the correlation between the classification score and localization accuracy.”
IoU-aware Single-stage Object Detector for Accurate Localization (2019, background): “PISA [18] assigns different weights to the positive examples in the classification loss based on their importance which is obtained by IoU Hierarchical Local Rank (IoU-HLR).”
IoU-balanced Loss Functions for Single-stage Object Detection (2019, methodology): “Normalization strategy [16] can also be used to keep the localization loss sum unchanged during the whole training procedure as Equ.”
IoU-uniform R-CNN: Breaking Through the Limitations of RPN (2019, methodology): “Different from existing NMS algorithms, Prime sample [3] investigates the sample importance and make the classifier more prone to give high scores to high IoU proposals.”
YOLO-FaceV2: A Scale and Occlusion Aware Face Detector (2022, methodology): “Prime Sample Attention (PISA) [22] proposed by Cao et al. assigns weights to positive and negative samples according to diﬀerent criteria.”
Imbalance Problems in Object Detection: A Review (2019, methodology): “CARL [30] is a method to promote examples with higher IoUs, and in such a way, 1.6% relative improvement is achieved compared to prime sampling without CARL (i.e. from 37.9 to 38.5).”
A Review on Deep Learning in UAV Remote Sensing (2021, background): “The Libra-RCNN [147], ATSS [7], Guided Anchoring [61], FSAF [216], PAA [145], GFL [65], PISA [153] and VFNet [191] detectors explore di ff erent sampling strategies and new loss metrics to improve the quality of selected positive samples and reduce the weight of the large negative samples.”
VarifocalNet: An IoU-aware Dense Object Detector (2020, unknown): “On the other hand, inspired by PISA [32] and IoU-balanced loss [33], we weight the positive example with the training target q .”
Learning From Noisy Anchors for One-Stage Object Detection (2019, background): “In addition, various hard [31, 26, 5, 24] and soft [22, 17, 3] sampling strategies have been proposed.”
Prime-Aware Adaptive Distillation (2020, background): “These samples are termed “prime samples” in a recent work [3].”