arvkr / bodyscan Goto Github PK

could you tell me what this error is about?
CUDA available: False
./data/inputs/./data/inputs/11.jpg,,,,,
[ WARN:[email protected]] global loadsave.cpp:248 findDecoder imread_('./data/inputs/./data/inputs/11.jpg'): can't open/read file: check file path/integrity
Traceback (most recent call last):
File "measure_body.py", line 174, in
depth_array = depth_predict(opt, image_pathfile='./data/depth.txt')
File "/home/amirreza/PycharmProjects/pythonProject/BodyScan/get_depth.py", line 13, in depth_predict
img_shape = [img.shape[0], img.shape[1]]
AttributeError: 'NoneType' object has no attribute 'shape'

why the path is ./data/inputs/./data/inputs/11.jpg,,,,,?

Hello, I was wondering if you can include performance metrics of the model? How do you measure accuracy, how does it compares to medical measurements?

Also is it possible to provide a few input pictures + results as body fat %?

Is there any demo website where we can test it?

When installing the needed libraries, I have an issue on this command, It was running forever:
(pytorch)$ pip3 install scikit-image opencv-python pandas h5py

The problem is when installing the opencv-python, and it can be fixed just by upgrading pip. One solution is to use pip install --upgrade pip before start the installation on the new virtual environment. So here is the new order of the commands:

(pytorch)$ pip install --upgrade pip
(pytorch)$ pip3 install torch torchvision
(pytorch)$ pip3 install scikit-image opencv-python pandas h5py
(pytorch)$ pip3 install cffi
(pytorch)$ pip3 install cython
(pytorch)$ pip3 install requests
(pytorch)$ pip3 install future

I get this error when building nms extension in google Colab. Any help?