Hung himself in the security field for many years, is a former ffdshow, xvid, ffmpeg, Live555 streamserver, DShow, D3D, OpenCV, and so on aspects of application development, used Grain - Media (GM8120, 8180), Maxim (Mobilygen MG3500), Amberalla (A2, A5S33 A5S55), TI DM365, DM368, Hisilicon Hi3518A, Hi3516C chips such as platform, the main development have IPCAM (web camera) and the DVR (DVR) products,

BBS just want to ask, do you have any friends in running Darknet + YOLOV3 training data set, and then transplanted to Hi3516CV500, 16 dv300 Hi3519AV100 NNIE, recognition effect how, high accuracy? Such as face, or do other object recognition, human recognition,,

we did haisi scheme before AI IPCAM (face captured, alignment machine) Hi3516AV200 + horizon, Hi3516CV500 + reading or FACE++ algorithm, Hi3520D, Hi3521A, Hi3531A, Hi3536D NVR, MDVR software, PC client software DXVA2, hard solution of CUDA, later want to do AI algorithms on Hisilicon platform, want to make technology like-minded friends, welcome to contact messages within the station,