#世界級
全球最神的 #物件偵測演算法 上線!
即時計算車流、為自駕車領航、計算社交距離……
🚗🚌🚐🚒🏎🚕🚌🚎🏍🚲🚐🚚
👌有了它,通通都搞定!
💁♂️ #YOLOv4 #最速最準 #免費開源釋出 #中研院與俄國開發者聯手打造
--
YOLO 是「You only look once」的簡稱,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,自2015年推出第一代版本後,深受開發者喜愛。
👍新一代的YOLOv4,由中研院資訊所廖弘源特聘研究員、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,平均正確率(Average Precision, AP)達43.5%,比前一代(YOLOv3)提高10%🎉
✅速度更快 ✅準確度更高 ✅成本低 使用一般的圖形處理器就能運算,「像是用算盤打出計算機的速度!」不需佈建雲端系統,即時分析回饋。
#YOLO可以拿來做什麼
廖弘源指出,團隊目前已運用該技術與義隆電子合作「智慧城市交通車流解決方案」,在路口就能進行交通影像辨識,例如即時分析車流、停等車列、車速,即時反映交通路況,調整號誌。目前已佈設於桃園、新竹。
#物件偵測的應用範圍不只在交通領域
YOLOv4已經於今年四月在Github(原始碼代管平台)公開,免費開源給全世界使用,各項應用層出不窮⬇⬇
👉計算社交距離
https://www.youtube.com/watch?v=obtERdHvM8o
👉偵測有無戴口罩
https://www.youtube.com/watch?v=yPTLc6FmQCM
✅這項技術的突破,也展現台灣在人工智慧研究領域擁有強大能量!了解更多開發過程:https://www.sinica.edu.tw/ch/news/6576
✅論文:https://arxiv.org/abs/2004.10934
#中研院 #資訊所 #YOLOv4 #objectdetection #computervision #影像辨識
_________________________________
Groundbreaking results in artificial intelligence achieved by international team with Taiwanese researchers! AS Distinguished Research Fellow Mark Liao and Postdoctoral Scholar Chien-Yao Wang from the Institute of Information Science worked with Alexey Bochkovskiy from Russia to develop YOLOv4, currently the fastest and most accurate object detection algorithm. YOLOv4 has an average precision (AP) rate of 43.5 percent, 10 percent higher than the previous version (YOLOv3).
Since the source code was shared on an open-source platform, tens of thousands of people around the world have tested YOLOv4 and developed their systems and products.
According to Mark Liao, YOLOv4 has been used in the development of “Smart City Traffic Flow Solutions”, a collaborative project with ELAN Microelectronics Corporation to enhance smart city innovation in Taiwan.
⏩ Press Release: https://www.sinica.edu.tw/en/news/6576
__________________________________
🎤媒體報導🎤
[自由時報] 人工智慧技術 分析車流準確率達9成
https://news.ltn.com.tw/news/life/paper/1383752
[蘋果日報] 交通指揮換AI!車少時一路綠燈 中研院研發最準車流辨識
https://reurl.cc/kdAbRK
[中央社] 台俄合作最速物件偵測 影像辨識解決交通車流
https://www.cna.com.tw/news/ait/202007020144.aspx
[警廣] 中研院演算法準確判斷車流 改善交通路況
https://reurl.cc/exYbeM
[公視] 中研院改良研發 世界級AI交通偵測系統
https://news.pts.org.tw/article/485383
[聯合報] 一眼揪出你有沒有超速!台俄聯手打造全球最神物件偵測
https://udn.com/news/story/7266/4673372
#塞車有解 #超速要當心 #AI都知道 🤫
同時也有10000部Youtube影片,追蹤數超過2,910的網紅コバにゃんチャンネル,也在其Youtube影片中提到,...
「object detection github」的推薦目錄:
- 關於object detection github 在 中央研究院 Academia Sinica Facebook 的最佳貼文
- 關於object detection github 在 AppWorks Facebook 的最佳解答
- 關於object detection github 在 コバにゃんチャンネル Youtube 的精選貼文
- 關於object detection github 在 大象中醫 Youtube 的精選貼文
- 關於object detection github 在 大象中醫 Youtube 的精選貼文
- 關於object detection github 在 Detectron2 is a platform for object detection, segmentation ... 的評價
object detection github 在 AppWorks Facebook 的最佳解答
【What's Up, AI? A Week of Tech Announcements】
This week in AI, Microsoft, Google, and Facebook all had some sort of announcement or press surrounding their AI activities. For all you founders interested in the world of AI, let's take a closer look together at them and what they've been up to recently.
Google: 4 new machine learning APIs for developers were released by Google earlier this week. The new APIs are (1) object detection, (2) tracking, (3) on-device translation, and (4) AutoML Vision Edge. This will help developers with more base APIs, UX, and greater customization. This release is on par with Google's vision for open source: bringing better technology to the world by promoting open source.
Facebook: On the topic of open source, Facebook continues with its commitment to the wider developer community (and specifically the AI community) by announcing Pythia, a modular plug-and-play framework that allows data scientists to quickly build, reproduce, and benchmark AI models. Also available on Github.
Microsoft: In an annual machine-translation competition (WMT19), Microsoft Research Asia came out top 8 of 19 different language-translation challenges, beating out other tech firms like Yandex and IBM. This is significant for them because they developed a speech-recognition system which transcribes as well as a human, so winning the competition solidifies their credibility in the space. Congrats Microsoft!
As we continue to see AI creation and adoption grow, we can only expect to hear about AI on the daily. In the meantime before AI becomes mainstream, check out this video from Code Bullet where he makes an AI that learns to walk while avoiding a death laser. Have a great weekend!
Applications are now open for AW#19. If you're an AI or blockchain startup, apply now here >> http://bit.ly/2JDsU8q
- Natalie Feng Lin, Analyst
object detection github 在 Detectron2 is a platform for object detection, segmentation ... 的必吃
Detectron2 is a platform for object detection, segmentation and other visual recognition tasks. - GitHub - facebookresearch/detectron2: Detectron2 is a ... ... <看更多>