中央研究院 資訊科學研究所

活動訊息

友善列印

學術演講

[資訊所/資創]前瞻科技演講系列_智慧交通系統與YOLOv4

  • 廖弘源、王建堯 博士 (中央研究院資訊科學研究所)
    邀請人:鐘楷閔、楊得年、蘇黎
  • 2020-09-15 (Tue.) 09:00 – 10:30
  • 資訊所新館106演講廳(現場);視訊連結請詳摘要

    會議號:170 593 6248
    密碼:yeJS6ubhp38
    https://asmeet.webex.com/asmeet/j.php?MTID=mf7dc354903e9abc9bffd1c6123a84996

    *此系列演講主要開放對象為本院資訊所及資創同仁

    **本所具與會者參加資格之認定,為確保演講品質,必要時得將解除與會權限(現場及視訊)

摘要

YOLOv4是一種利用人工智慧執行即時物件偵測(Object Detection)的技術,能偵測物件、追蹤及判斷,可應用於交通車流計算、自駕車研發、工廠瑕疵檢測、醫療影像分析、五官定位等。自今(109)年4月以開放原始碼免費釋出後,全世界已有數萬人測試應用;研究團隊將說明此物件偵測演算法如何識別圖片或影像裡的物件。

過去訓練電腦識別一張照片裡的物件,須執行幾千次的識別指令。YOLO是「You Only Look Once」的簡稱,顧名思義,只需訓練一個網路模型,電腦只要看一眼,就能判斷照片或影像裡的物件類別與位置,大大提升辨識速度。自2015年推出第一代版本後,深受開發者喜愛,為電腦視覺(Computer vision)技術立下重要的里程碑。改良後的YOLOv4性能更強大,經微軟開源影像資料庫(MSCOCO)測試,辨識物件的速度及精確度皆大幅提升,堪稱是目前世界上辨識速度最快、最精準的物件偵測演算法。

廖弘源特聘研究員、王建堯博士後研究員,與俄羅斯開發者博科夫斯基(Alexey Bochkovskiy)共同研發出目前世界上最快最準的物件偵測演算法(YOLOv4),平均正確率(Average Precision, AP)達43.5%,比前一代(YOLOv3)提高10%,更一舉超越其他種影像辨識技術。同時也運用該技術與義隆電子合作開發「智慧城市交通車流解決方案」,目前已佈設於桃園、新竹,在路口就能進行交通影像辨識及車流分析,為我國智慧城市發展往前邁出一大步。

BIO

Mark Liaoreceived his Ph.D degree in electrical engineering from Northwestern University in 1990. In July 1991, he joined the Institute of Information Science, Academia Sinica, Taiwan and currently, is a Distinguished Research Fellow and the Director. He has worked in the fields of multimedia signal processing, image processing, computer vision, pattern recognition, video forensics, and multimedia protection for more than 25 years.  Since November 2016, he has been appointed a Chair Professor of National Chiao-Tung University. He received the ''Young Investigators'' Award from Academia Sinica in 1998; the Distinguished Research Award from the National Science Council of Taiwan in 2003, 2010 and 2013; and the Academia Sinica Investigator Award in 2010. In 2016, he received the TECO award from the TECO foundation. His professional activities include: Co-Chair, 2004 International Conference on Multimedia and Exposition (ICME); Technical Co-chair, 2007 ICME; President, Image Processing and Pattern Recognition Society of Taiwan (2006-08); Editorial Board Member, IEEE Signal Processing Magazine (2010-13); Associate Editor, IEEE Transactions on Image Processing (2009-13), IEEE Transactions on Information Forensics and Security (2009-12) and IEEE Transactions on Multimedia (1998-2001).  He has been a Fellow of the IEEE since 2013 for contributions to image and video forensics and security. Dr. Liao currently serves as an AE of ACM Computing Surveys.

 

 

Chien-Yao Wangis currently a postdoc fellow with the Institute of Information Science, Academia Sinica, Taiwan. He received the B.S. degree in computer science and information engineering from National Central University, Zhongli, Taiwan, in 2013, and he received the Ph.D. degree from National Central University, Zhongli, Taiwan, in 2017. His research interests include signal processing, deep learning, and machine learning. He is an honorary member of Phi Tau Phi Scholastic Honor Society.