AutoContrast in Albumentations?

Can I use AutoContrast in Albumentations library? That is, making the darkest pixel black and lightest pixel white. I see the RandomBrightnessContrast one, but is it the right one when I select contrast parameter to 0?

Unfortunately, there is no ready-to-use transform for autocontrast now. You may implement it by yourself using Lambda transform:

def _autocontrast(img):
    h = cv2.calcHist([img], [0], None, [256], (0, 256)).ravel()

    for lo in range(256):
        if h[lo]:
            break
    for hi in range(255, -1, -1):
        if h[hi]:
            break

    if hi > lo:
        lut = np.zeros(256, dtype=np.uint8)
        scale_coef = 255.0 / (hi - lo)
        offset = -lo * scale_coef
        for ix in range(256):
            lut[ix] = int(np.clip(ix * scale_coef + offset, 0, 255))

        img = cv2.LUT(img, lut)

    return img

transform = Lambda(image = lambda img: _autocontrast(img))

P.S. Pull request containing Autocontrast transform is under review for a long time. You may vote for this PR on github in order to highlight it to maintainers :)