An AdaBoost object detection design for heterogeneous computing with OpenCL

Bing Yang Cheng; Jui Sheng Lee; Jiun-In  Guo

doi:10.1109/ICCE-TW.2015.7216901

An AdaBoost object detection design for heterogeneous computing with OpenCL

Bing Yang Cheng, Jui Sheng Lee, Jiun-In Guo

電子研究所

研究成果: Conference contribution › 同行評審

1 引文斯高帕斯（Scopus）

摘要

AdaBoost classification with Haar-like features [1] is commonly adopted for object detection. Feature calculation in AdaBoost algorithm is the most time-consuming part, which occupies over 98% of the computation and cannot reach realtime processing with CPU computing only. In this paper we propose an object detection design for heterogeneous computing with OpenCL. By adopting the techniques of scale parallelizing, stage partitioning, and dynamic stage scheduling on AdaBoost algorithm, the proposed design solves load-unbalanced problems when realize in multicore CPU and GPU platform. The proposed object detection design achieves 32.5 fps at D1 resolution on an AMD A10-7850K processor.

原文	English
主出版物標題	2015 IEEE International Conference on Consumer Electronics - Taiwan, ICCE-TW 2015
發行者	Institute of Electrical and Electronics Engineers Inc.
頁面	286-287
頁數	2
ISBN（電子）	9781479987443
DOIs	https://doi.org/10.1109/ICCE-TW.2015.7216901
出版狀態	Published - 20 8月 2015
事件	2nd IEEE International Conference on Consumer Electronics - Taiwan, ICCE-TW 2015 - Taipei, 台灣持續時間: 6 6月 2015 → 8 6月 2015

出版系列

名字	2015 IEEE International Conference on Consumer Electronics - Taiwan, ICCE-TW 2015

Conference

Conference	2nd IEEE International Conference on Consumer Electronics - Taiwan, ICCE-TW 2015
國家/地區	台灣
城市	Taipei
期間	6/06/15 → 8/06/15

存取文件

10.1109/ICCE-TW.2015.7216901

其它文件與鏈接

Link to publication in Scopus

引用此

Cheng, B. Y., Lee, J. S., & Guo, J.-I. (2015). An AdaBoost object detection design for heterogeneous computing with OpenCL. 於 2015 IEEE International Conference on Consumer Electronics - Taiwan, ICCE-TW 2015 (頁 286-287). 文章 7216901 (2015 IEEE International Conference on Consumer Electronics - Taiwan, ICCE-TW 2015). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICCE-TW.2015.7216901

Cheng, Bing Yang ; Lee, Jui Sheng ; Guo, Jiun-In . / An AdaBoost object detection design for heterogeneous computing with OpenCL. 2015 IEEE International Conference on Consumer Electronics - Taiwan, ICCE-TW 2015. Institute of Electrical and Electronics Engineers Inc., 2015. 頁 286-287 (2015 IEEE International Conference on Consumer Electronics - Taiwan, ICCE-TW 2015).

@inproceedings{de51f3f0c09a43289c23fa882be38831,

title = "An AdaBoost object detection design for heterogeneous computing with OpenCL",

abstract = "AdaBoost classification with Haar-like features [1] is commonly adopted for object detection. Feature calculation in AdaBoost algorithm is the most time-consuming part, which occupies over 98% of the computation and cannot reach realtime processing with CPU computing only. In this paper we propose an object detection design for heterogeneous computing with OpenCL. By adopting the techniques of scale parallelizing, stage partitioning, and dynamic stage scheduling on AdaBoost algorithm, the proposed design solves load-unbalanced problems when realize in multicore CPU and GPU platform. The proposed object detection design achieves 32.5 fps at D1 resolution on an AMD A10-7850K processor.",

keywords = "Algorithm design and analysis, Central Processing Unit, Dynamic scheduling, Face detection, Graphics processing units, Heuristic algorithms, Object detection",

author = "Cheng, {Bing Yang} and Lee, {Jui Sheng} and Jiun-In Guo",

year = "2015",

month = aug,

day = "20",

doi = "10.1109/ICCE-TW.2015.7216901",

language = "English",

series = "2015 IEEE International Conference on Consumer Electronics - Taiwan, ICCE-TW 2015",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "286--287",

booktitle = "2015 IEEE International Conference on Consumer Electronics - Taiwan, ICCE-TW 2015",

address = "美國",

note = "2nd IEEE International Conference on Consumer Electronics - Taiwan, ICCE-TW 2015 ; Conference date: 06-06-2015 Through 08-06-2015",

}

Cheng, BY, Lee, JS & Guo, J-I 2015, An AdaBoost object detection design for heterogeneous computing with OpenCL. 於 2015 IEEE International Conference on Consumer Electronics - Taiwan, ICCE-TW 2015., 7216901, 2015 IEEE International Conference on Consumer Electronics - Taiwan, ICCE-TW 2015, Institute of Electrical and Electronics Engineers Inc., 頁 286-287, 2nd IEEE International Conference on Consumer Electronics - Taiwan, ICCE-TW 2015, Taipei, 台灣, 6/06/15. https://doi.org/10.1109/ICCE-TW.2015.7216901

An AdaBoost object detection design for heterogeneous computing with OpenCL. / Cheng, Bing Yang; Lee, Jui Sheng; Guo, Jiun-In .
2015 IEEE International Conference on Consumer Electronics - Taiwan, ICCE-TW 2015. Institute of Electrical and Electronics Engineers Inc., 2015. p. 286-287 7216901 (2015 IEEE International Conference on Consumer Electronics - Taiwan, ICCE-TW 2015).

研究成果: Conference contribution › 同行評審

TY - GEN

T1 - An AdaBoost object detection design for heterogeneous computing with OpenCL

AU - Cheng, Bing Yang

AU - Lee, Jui Sheng

AU - Guo, Jiun-In

PY - 2015/8/20

Y1 - 2015/8/20

N2 - AdaBoost classification with Haar-like features [1] is commonly adopted for object detection. Feature calculation in AdaBoost algorithm is the most time-consuming part, which occupies over 98% of the computation and cannot reach realtime processing with CPU computing only. In this paper we propose an object detection design for heterogeneous computing with OpenCL. By adopting the techniques of scale parallelizing, stage partitioning, and dynamic stage scheduling on AdaBoost algorithm, the proposed design solves load-unbalanced problems when realize in multicore CPU and GPU platform. The proposed object detection design achieves 32.5 fps at D1 resolution on an AMD A10-7850K processor.

AB - AdaBoost classification with Haar-like features [1] is commonly adopted for object detection. Feature calculation in AdaBoost algorithm is the most time-consuming part, which occupies over 98% of the computation and cannot reach realtime processing with CPU computing only. In this paper we propose an object detection design for heterogeneous computing with OpenCL. By adopting the techniques of scale parallelizing, stage partitioning, and dynamic stage scheduling on AdaBoost algorithm, the proposed design solves load-unbalanced problems when realize in multicore CPU and GPU platform. The proposed object detection design achieves 32.5 fps at D1 resolution on an AMD A10-7850K processor.

KW - Algorithm design and analysis

KW - Central Processing Unit

KW - Dynamic scheduling

KW - Face detection

KW - Graphics processing units

KW - Heuristic algorithms

KW - Object detection

UR - http://www.scopus.com/inward/record.url?scp=84959484257&partnerID=8YFLogxK

U2 - 10.1109/ICCE-TW.2015.7216901

DO - 10.1109/ICCE-TW.2015.7216901

M3 - Conference contribution

AN - SCOPUS:84959484257

T3 - 2015 IEEE International Conference on Consumer Electronics - Taiwan, ICCE-TW 2015

SP - 286

EP - 287

BT - 2015 IEEE International Conference on Consumer Electronics - Taiwan, ICCE-TW 2015

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 2nd IEEE International Conference on Consumer Electronics - Taiwan, ICCE-TW 2015

Y2 - 6 June 2015 through 8 June 2015

ER -

Cheng BY, Lee JS, Guo JI. An AdaBoost object detection design for heterogeneous computing with OpenCL. 於 2015 IEEE International Conference on Consumer Electronics - Taiwan, ICCE-TW 2015. Institute of Electrical and Electronics Engineers Inc. 2015. p. 286-287. 7216901. (2015 IEEE International Conference on Consumer Electronics - Taiwan, ICCE-TW 2015). doi: 10.1109/ICCE-TW.2015.7216901

An AdaBoost object detection design for heterogeneous computing with OpenCL

摘要

出版系列

Conference

存取文件

其它文件與鏈接

指紋

引用此