Can DeepFace verify() accept an image array or PIL Image object?

user_16019935 0
Viewed 575

My DeepFace Implementation

def verify(img, frame, model):
    results= DeepFace.verify(img, frame, enforce_detection=False, model_name=model)
    print("result: ", results)
    verification= results['verified']
    if verification is True:
        print(model, "Successfully recognised. SCORE=")
        return
    return

I am trying to pass a PIL Image instead of "*.jpg" to verify() function but it gives error

    raise ValueError("Invalid arguments passed to verify function: ", instance)
ValueError: ('Invalid arguments passed to verify function: ', <PIL.Image.Image image mode=RGB size=160x160 at 0x7F4D5D03FBE0>)

I want to ask is there any way I can pass directly PIL Image object or image array without saving it to disk prior?

What I am passing:

picture= "extracted_face_picture/single_face_picture.jpg"
picture= Image.open(picture)
picture= picture.resize((160,160))

frames_from_npz= "video_faces.npz"
frames= np.load(frames_from_npz)
frames= frames["arr_0"]
i=0
for frames_arr in frames:
    frame= Image.fromarray(frames_arr)

    df.verify(picture,frame, "Facenet")
    print(i, "Above reuslts are for frame", frame_num)
    
    i+= 1

Note that I can still successfully implement DeepFace with saving images in directories then reading them with imread() I only want to know if there are other ways without saving images to disk

verify python-imaging-library image python
2 Answers

If you are using this module, the documentation says:

Herein, face pairs could be exact image paths, numpy array or base64 encoded images

So, presumably, you can make your PIL Images into Numpy arrays like this:

results = DeepFace.verify(np.array(PILIMAGE), ...)
user_28366210

Thanks a lot brother, I followed your way and got my results.

For others: I encountered error again once I followed brother Mark Setchell but it was precisely because I was passing one input as path and other one as numpy array. If you have to pass numpy array make sure to pass it for both of the input parameters.

picture= "extracted_face_picture/single_face_picture.jpg"
picture= Image.open(picture)
.
.
df.verify(picture, np.array(frame), "Facenet")

correction:

df.verify(np.array(picture),np.array(frame), "Facenet")

Thanks a lot brother Mark Setchell.

user_160199350
Related