Work-Efficient Parallel Non-Maximum Suppression Kernels
Authors:
- 21 August 2020
- journal article
- research article
- Published by Oxford University Press (OUP) in The Computer Journal
- Vol. 65 (4) , 773-787
Abstract
In the context of object detection, sliding-window classifiers and single-shot convolutional neural network (CNN) meta-architectures typically yield multiple overlapping candidate windows with similar high scores around the true location of a particular object. Non-maximum suppression (NMS) is the process of selecting a single representative candidate within this cluster of detections, so as to obtain a unique detection per object appearing on a given picture. In this paper, we present a highly scalable NMS algorithm for embedded graphics processing unit (GPU) architectures that is designed from scratch to handle workloads featuring thousands of simultaneous detections on a given picture. Our kernels are directly applicable to other sequential NMS algorithms such as FeatureNMS, Soft-NMS or AdaptiveNMS that share the inner workings of the classic greedy NMS method. The obtained performance results show that our parallel NMS algorithm is capable of clustering 1024 simultaneous detected objects per frame in roughly 1 ms on both Tegra X1 and Tegra X2 on-die GPUs, while taking 2 ms on Tegra K1. Furthermore, our proposed parallel greedy NMS algorithm yields a 14–40x speed up when compared to state-of-the-art NMS methods that require learning a CNN from annotated data.Keywords
All Related Versions
Funding Information
- Ministerio de Economía y Competitividad (TIN2015-65316-P, TEC2012-38939-C03-02)
- Departament d’Innovacio, Universitats i Empresa de la Generalitat de Catalunya (2014-SGR-1051)
- European Commission (H2020-ICT-644312)
This publication has 23 references indexed in Scilit:
- Work-efficient parallel non-maximum suppression for embedded GPU architectures2016
- Multi-view Face Detection Using Deep Convolutional Neural Networks2015
- A convolutional neural network cascade for face detection2015
- Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation2014
- The Pascal Visual Object Classes (VOC) ChallengeInternational Journal of Computer Vision, 2009
- Mean Shift Parallel Tracking on GPU2009
- Efficient K-Means Clustering Using Accelerated Graphics Processors2008
- The Boosting Approach to Machine Learning: An Overview2003
- Support-vector networksMachine Learning, 1995
- Parallelism in random access machines1978