Encoding is not prepared for certain images

[nidhiSingh89]

• face_recognition version: 1.3.0
• Python version: 3.7
• Operating System: ubuntu

Description

I am using this piece of code for generating encoding of the images for training the model.

    image = PIL.Image.open(services.utils.get_image_from_url(image_url))
    image = ImageOps.exif_transpose(image)
    im_gray = np.array(image)
    rgb = cv.cvtColor(im_gray, cv.COLOR_BGR2RGB)
    boxes = face_recognition.face_locations(rgb,model='hog')
    encodings = face_recognition.face_encodings(rgb, boxes)

This code is working fine for most of the images but for certain images this is generating boxes and encodings as empty array.

What I Did

I tried using cnn model along with number_of_times_to_upsample = 1, but no success. Please help in identifying the issue. I am attaching the images for reference in which we are facing this issue.

[Gaotr2022]

Maybe the algorithm didn't detect face in these pictures.

[nidhiSingh89]

Maybe the algorithm didn't detect face in these pictures.

Yeah I understand that, but unable to find the reason for the same. As pictures are clicked directly from the camera and uploaded.
Can you please help in identifying the reason for it.