Analyze with numpy input.

[PankajSavaliya]

img = BytesIO(requests.get(image.url).content)
frame_byte = bytearray(urllib.request.urlopen(data.image.url).read())
np_array = numpy.asarray(frame_byte, dtype=numpy.uint8)
frame = cv2.imdecode(np_array,
                     cv2.IMREAD_UNCHANGED)  # 'Load it as it is'

print(np_array, type(np_array))  # type is numpy.ndarray
print(frame, type(frame))  # type is numpy.ndarray

obj = DeepFace.analyze(img_path=np_array, actions=['age', 'gender', ],
                       enforce_detection=False)

both output types are numpy.ndarray and np_array getting an error and the second one getting the right result, can you please explain why the first one is not working even it's the right type for method?.

[FranSPG]

If i reproduced exactly the error you got, I think that the problem here is even though they are both NumPy ndarray, their shapes are not equals, you can see this with np_array.shape and frame.shape.

You need to process the binary as an image and that's the job that you're doing in the line

frame = cv2.imdecode(np_array, cv2.IMREAD_UNCHANGED) # 'Load it as it is'.

If this doesn't solve your error, please provide more information about the error and maybe I could help you.