Errors when running python3 demo.py

[LC9]

I have run python3 demo.py code and I got this /usr/lib/python3.5/importlib/_bootstrap.py:222: RuntimeWarning: numpy.dtype size changed, may indicate binary incompatibility. Expected 96, got 88 return f(*args, **kwds) Using TensorFlow backend. /usr/lib/python3.5/importlib/_bootstrap.py:222: RuntimeWarning: numpy.dtype size changed, may indicate binary incompatibility. Expected 96, got 88 return f(*args, **kwds) /usr/lib/python3.5/importlib/_bootstrap.py:222: RuntimeWarning: numpy.dtype size changed, may indicate binary incompatibility. Expected 96, got 88 return f(*args, **kwds) Downloading data from https://github.com/yu4u/age-gender-estimation/releases/download/v0.5/weights.18-4.06.hdf5 195813376/195810464 [==============================] - 56s 0us/step 2018-07-24 16:57:16.697715: I tensorflow/core/platform/cpu_feature_guard.cc:141] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA VIDEOIO ERROR: V4L: can't open camera by index 0 Traceback (most recent call last): File "demo.py", line 121, in <module> main() File "demo.py", line 82, in main for img in yield_images(): File "demo.py", line 60, in yield_images yield img File "/usr/lib/python3.5/contextlib.py", line 77, in __exit__ self.gen.throw(type, value, traceback) File "demo.py", line 42, in video_capture yield cap File "demo.py", line 58, in yield_images raise RuntimeError("Failed to capture image") RuntimeError: Failed to capture image

NOTE: I'm on Ubuntu 16.04 installed in VMware and I have a webcam in my laptop, I think I've installed all the dependencies correctly

[yu4u]

Failed in capturing an image from the webcam. https://github.com/yu4u/age-gender-estimation/blob/master/demo.py#L57

The webcam may not be recognized from the virtual machine. Did you allow the virtual machine to used the webcam (as a USB device?) ?

[LC9]

I don’t think so, but my webcam is built-in. Do you know how can I check if the webcam is used as a USB device or how can I allow the VM to use it as USB device

[yu4u]

As I could not reproduce your environment, I have no idea. I found many related article by searching with "vmware built in webcam".

[LC9]

Thanks alot you are right, I haven't allow the webcam to work in VM, I have managed to make it work, but after the webcam started do I have to stop it manually ? I mean stop the program demo.py ?

[yu4u]

Usually there is no need to start/stop a webcam after VM recognized a webcam but I'm not sure there is any VM-specific problem or not because I have never used a webcam from VMs.

[LC9]

I mean after running this python3 demo.py, the webcam started and it get my age and gender changing then what next ? should I close the window or wait ?

[yu4u]

demo.py run in infinite while loop. Pressing ESC key would finish the program.

[LC9]

Yeah, I did that thnx a lot :) I have another question, can I analyze one image by itself using ur algorithm, in order to analyze my dataset?

[yu4u]

for img in yield_images():

Instead of the above line in demo.py, you can read an image from file:

img = cv2.imread(image_path)

or you can loop:

from pathlib import Path

for image_path in Path(path_to_img_dir).glob("*.*"):
    img = cv2.imread(str(image_path))

[LC9]

So you mean I should commented out for img in yield_images(): and put this code instead img = cv2.imread(image_path) image_path means the path of the image and its format like JPG, TIF, PNG etc. ?

and If I want to analyze my own dataset of 5000 images ? how can I do it and is it possible to do it without labeling them ?

I'm asking those questions coz I'm doing a research on this topic

Thnx

[yu4u]

and If I want to analyze my own dataset of 5000 images ?

As described above, you can iterate over all files in path_to_img_dir directory by:

from pathlib import Path

for image_path in Path(path_to_img_dir).glob("*.*"):
    img = cv2.imread(str(image_path))

[LC9]

File "demo.py", line 83 img = cv2.imread('/home/nasser/2.jpeg') ^ IndentationError: unexpected indent

I tried to analyze this picture with no avail, what do you think is wrong ?

I did like what you said check: ` #load model and weights img_size = 64 model = WideResNet(img_size, depth=depth, k=k)() model.load_weights(weight_file)

##for img in yield_images():
 img = cv2.imread('/home/nasser/2.jpeg')    
    input_img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
    img_h, img_w, _ = np.shape(input_img)`

[LC9]

I tried to do 1 image but I get this error 'line 83 img = cv2.imread('/home/nasser/2.jpeg') ^ IndentationError: unexpected indent`

when I tried to iterate a directory of images: File "demo.py", line 84 for image_path in Path(“/home/nasser/google-images-download/Family/dataset_3000”).glob("*.*"): ^ IndentationError: unexpected indent check the screeenshot:

[yu4u]

Both errors clearly indicate what's wrong; indent.

[LC9]

Thnx I managed to fix it, but after like 300-400 pictures it stopped and gave me this error: File "demo.py", line 124, in main() File "demo.py", line 86, in main input_img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB) cv2.error: OpenCV(3.4.2) /io/opencv/modules/imgproc/src/color.hpp:253: error: (-215:Assertion failed) VScn::contains(scn) && VDcn::contains(dcn) && VDepth::contains(depth) in function 'CvtHelper'

[yu4u]

Would you kindly solve by yourself problems not specific to this project. In this case, the following modification may solve the problem...

img = cv2.imread(str(image_path), 1)

[LC9]

Hello @yu4u I have a question related to this project, when change add the code you mention above to analyze one picture, I recognize that ages range from 3-14 in pictures are not detected, for example if I put a baby face and his age range from 3-5 the demo.py can’t detect the face and didn’t give results, I tried many pictures and age ranges from 3-14.

Do you have any suggestion ?

Thnx a lot

[yu4u]

It depends on the accuracy of face detector. Please try dlib.cnn_face_detection_model_v1 detector instead of current dlib.get_frontal_face_detector().

[LC9]

I have change the detector you provided instead

# for face detection detector = dlib.cnn_face_detection_model_v1

# load model and weights img_size = 64 model = WideResNet(img_size, depth=depth, k=k)() model.load_weights(weight_file)

then I get this error

Traceback (most recent call last): File "demo.py", line 124, in <module> main() File "demo.py", line 90, in main detected = detector(input_img, 1) TypeError: __init__(): incompatible constructor arguments. The following argument types are supported: 1. dlib.cnn_face_detection_model_v1(filename: str)

Invoked with: array([[[180, 163, 137], [179, 162, 136], [178, 161, 135], ..., [129, 118, 100], [129, 118, 100], [129, 118, 100]], [[179, 162, 136], [179, 162, 136], [179, 162, 136], ..., [129, 118, 100], [129, 118, 100], [129, 118, 100]], [[178, 161, 135], [179, 162, 136], [179, 162, 136], ..., [129, 118, 100], [129, 118, 100], [129, 118, 100]], ..., [[155, 166, 184], [153, 164, 182], [152, 163, 181], ..., [119, 130, 148], [120, 131, 149], [121, 132, 150]], [[155, 166, 184], [152, 163, 181], [152, 163, 181], ..., [118, 129, 147], [118, 129, 147], [118, 129, 147]], [[153, 164, 182], [151, 162, 180], [150, 161, 179], ..., [119, 130, 148], [117, 128, 146], [117, 128, 146]]], dtype=uint8), 1>

I am sorry but I think this problem related to the project as well, is there something missing ?

[yu4u]

It's simply the problem of how to use dlib CNN detector (dlib.cnn_face_detection_model_v1). Please refer to dlib documents.

[anmolpassi092]

I have run python3 demo.py code and I got this

File "demo.py", line 138, in main() File "demo.py", line 87, in main file_hash=modhash, cache_dir=Path(file).resolve().parent) File "/home/anmol/.local/lib/python3.5/site-packages/keras/utils/data_utils.py", line 177, in get_file datadir_base = os.path.expanduser(cache_dir) File "/usr/lib/python3.5/posixpath.py", line 228, in expanduser if not path.startswith(tilde): AttributeError: 'PosixPath' object has no attribute 'startswith'

[yu4u]

Could you try to change

file_hash=modhash, cache_dir=Path(__file__).resolve().parent)

to

file_hash=modhash, cache_dir=str(Path(__file__).resolve().parent))

?

[anmolpassi092]

Yes, It worked...thanks