Bad shape for unary energy (Need (2, 1053979828), got (39677, 26564))

[pratik-segmind]

My Code:

def mask_smoothen(img, mask, nclasses):

    mask = mask.astype(np.float32) #mask.astype(np.uint8)
    image = img
    height, width, channels = image.shape

    d = dcrf.DenseCRF2D(height, width, nclasses)

    labels = mask.T

    U = unary_from_softmax(labels)
    U = np.ascontiguousarray(U) 
    print(U.dtype, U.shape)

    d.setUnaryEnergy(U)

    d.addPairwiseBilateral(sxy=80, srgb=13, rgbim=image, compat=10)

    Q = d.inference(5)

    fine_mask = np.argmax(Q, axis=0).reshape(height, width)

    return fine_mask

I get the error "Bad shape for unary energy (Need (2, 1053979828), got (39677, 26564))"

My image and masks are large: 39677 x 26564

Can some one help me understand what is happening here.

[andres-fr]

The readme says that you need to flatten it to the shape (NUM_CLASSES, H*W). Hope that helps!

[huixiancheng]

resize you pic maybe