quanhua92 / human-pose-estimation-opencv Goto Github PK

Is there an implementation on using the IE Model which has .xml and .bin models for pose estimation?

hey,
have you tried running this on an embedded platform like a raspberry pi etc??
if yes, how was the accuracy and inference time??

only single person,not multiperson

Hello, I was testing the project, it is so good, but the problema is that when I want to detect the keyt points of half body it does not detect, specially when it is legs, like this image:

Is there a way that it can detect the poses in this kind o images?
Thank you so much

error: (-2:Unspecified error) FAILED: fs.is_open(). Can't open "graph_opt.pb" in function

I tested your model in pictures of multiple people and found that it seems that it is not possible to detect multiple people at the same time. Then I found a annotation in your program 'Originally, we try to find all the local maximums. To simplify a sample we just find a global one. However, only a single pose at the same time can be detected this way.' So is this model suitable for multi-person pose detection?If it can, how can I improve it?

I have tried this model on the google colab.

!python openpose.py --input image5.jpg
:cannot connect to X server

please suggest the solution.

I tried using the other graphs in place of the mobilenet_thin but they don't work. Any idea on how to get more precision?

https://timgsa.baidu.com/timg?image&quality=80&size=b9999_10000&sec=1544691320461&di=3a212dc919f9e02968e26704a2040ea2&imgtype=0&src=http%3A%2F%2Fimgsrc.baidu.com%2Fimgad%2Fpic%2Fitem%2F7acb0a46f21fbe0909572e3d60600c338744ad8a.jpg

Hi, i'm interested in training MobileNet openpose, could you give me some advices? Thanks for your help.

因为正在做毕设想像您学习一下，有偿也行，想借鉴一下训练的代码！谢谢！

The detection of finger joint in gray image seems to be incorrect.

Hi,
My question is: For a single image, we get the output of dimension (1194747) where what 1 and 19 represent is understood. But for a single image, what does the 4747 matrix represents? Please reply!!

具体应该怎么样从读取摄像头改为读取图片啊，求救

I am getting below output on test image

当我使用 ildoonet/tf-pose-estimation
中提供的mobilenet_v2_small 模型时，有如下错误

File "openpose.py", line 29, in
net = cv.dnn.readNetFromTensorflow("graph_opt.pb")
cv2.error: OpenCV(3.4.4) /io/opencv/modules/dnn/src/tensorflow/tf_importer.cpp:535: error: (-2:Unspecified error) Input [strided_slice] for node [base/layer_14/output/upsample] not found in function 'getConstBlob'

请问这个是什么原因呢，谢谢！！！