KwanHua Lee

Sorry for the very late reply, you could get your camera and distortion matrix through calibration. Code below would help. https://github.com/opencv/opencv/tree/master/samples/cpp/tutorial_code/calib3d/camera_calibration

Hi, Sorry for the very late reply. Can you compile now? It relies on OpenCV and dlib. You can find how to install them below: https://docs.opencv.org/2.4/doc/tutorials/introduction/linux_install/linux_install.html http://dlib.net/compile.html

It's very weird that your solvePnP() running so slow. I have tested on Quad-core ARM® Cortex® A53 with 1.2 GHz per core, and the result is **1.5ms**. Do you build...

It depends on 2D facial landmark localization accuracy. Maybe you could try to use mtcnn to replace dlib's method.

@ASCharan Is there any error message?

你好，我覺得應該是內參數的問題，你可以先用opencv自帶的程式校正一下K和D兩個參數 sample code在(opencv/samples/cpp/calibration.cpp) 另外請問你的video input解析度是多少呢，我的是640*480

多加一點2D-3D點的連結對於最後逼近解會較準確，目前只用了14個點

I'm sorry its a hard work and takes time to implement from scratch.

Hi @yacaeh I add darknet library path to LD_LIBRARY_PATH. vi ~/.bashrc #put the following at the bottom export LD_LIBRARY_PATH=path_to_your_darknet_folder"${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}} source ~/.bashrc

Hi @yacaeh Looks like something wrong in netout size (Line 560 in main.cpp) 1. Have you change the net width and height in openpose.cfg? You could check if the final...