[fbsync] [proto] Speed improvements for adjust hue op (#6805)

Summary:
* WIP

* Updated rgb2hsv and a bit of hsv2rgb

* Fix issue with batch of images

* Few improvements

* hsv2rgb improvements

* PR review

* another update

* Fix cuda issue with empty images
torch.aminmax is failing

Reviewed By: YosuaMichael

Differential Revision: D40722899

fbshipit-source-id: 59edbba970a015fbc58c26828b36197945f46080

Co-authored-by: Vasilis Vryniotis <[email protected]>

Author: datumbox

torchvision/prototype/transforms/functional/_color.py

@@ -143,7 +143,104 @@ def adjust_sharpness(inpt: features.InputTypeJIT, sharpness_factor: float) -> fe
         return adjust_sharpness_image_pil(inpt, sharpness_factor=sharpness_factor)
 
 
-adjust_hue_image_tensor = _FT.adjust_hue
+def _rgb_to_hsv(image: torch.Tensor) -> torch.Tensor:
+    r, g, _ = image.unbind(dim=-3)
+
+    # Implementation is based on
+    # https://github.com/python-pillow/Pillow/blob/4174d4267616897df3746d315d5a2d0f82c656ee/src/libImaging/Convert.c#L330
+    minc, maxc = torch.aminmax(image, dim=-3)
+
+    # The algorithm erases S and H channel where `maxc = minc`. This avoids NaN
+    # from happening in the results, because
+    #   + S channel has division by `maxc`, which is zero only if `maxc = minc`
+    #   + H channel has division by `(maxc - minc)`.
+    #
+    # Instead of overwriting NaN afterwards, we just prevent it from occuring so
+    # we don't need to deal with it in case we save the NaN in a buffer in
+    # backprop, if it is ever supported, but it doesn't hurt to do so.
+    eqc = maxc == minc
+
+    channels_range = maxc - minc
+    # Since `eqc => channels_range = 0`, replacing denominator with 1 when `eqc` is fine.
+    ones = torch.ones_like(maxc)
+    s = channels_range / torch.where(eqc, ones, maxc)
+    # Note that `eqc => maxc = minc = r = g = b`. So the following calculation
+    # of `h` would reduce to `bc - gc + 2 + rc - bc + 4 + rc - bc = 6` so it
+    # would not matter what values `rc`, `gc`, and `bc` have here, and thus
+    # replacing denominator with 1 when `eqc` is fine.
+    channels_range_divisor = torch.where(eqc, ones, channels_range).unsqueeze_(dim=-3)
+    rc, gc, bc = ((maxc.unsqueeze(dim=-3) - image) / channels_range_divisor).unbind(dim=-3)
+
+    mask_maxc_neq_r = maxc != r
+    mask_maxc_eq_g = maxc == g
+    mask_maxc_neq_g = ~mask_maxc_eq_g
+
+    hr = (bc - gc).mul_(~mask_maxc_neq_r)
+    hg = (2.0 + rc).sub_(bc).mul_(mask_maxc_eq_g & mask_maxc_neq_r)
+    hb = (4.0 + gc).sub_(rc).mul_(mask_maxc_neq_g & mask_maxc_neq_r)
+
+    h = hr.add_(hg).add_(hb)
+    h = h.div_(6.0).add_(1.0).fmod_(1.0)
+    return torch.stack((h, s, maxc), dim=-3)
+
+
+def _hsv_to_rgb(img: torch.Tensor) -> torch.Tensor:
+    h, s, v = img.unbind(dim=-3)
+    h6 = h * 6
+    i = torch.floor(h6)
+    f = (h6) - i
+    i = i.to(dtype=torch.int32)
+
+    p = (v * (1.0 - s)).clamp_(0.0, 1.0)
+    q = (v * (1.0 - s * f)).clamp_(0.0, 1.0)
+    t = (v * (1.0 - s * (1.0 - f))).clamp_(0.0, 1.0)
+    i.remainder_(6)
+
+    mask = i.unsqueeze(dim=-3) == torch.arange(6, device=i.device).view(-1, 1, 1)
+
+    a1 = torch.stack((v, q, p, p, t, v), dim=-3)
+    a2 = torch.stack((t, v, v, q, p, p), dim=-3)
+    a3 = torch.stack((p, p, t, v, v, q), dim=-3)
+    a4 = torch.stack((a1, a2, a3), dim=-4)
+
+    return (a4.mul_(mask.to(dtype=img.dtype).unsqueeze(dim=-4))).sum(dim=-3)
+
+
+def adjust_hue_image_tensor(image: torch.Tensor, hue_factor: float) -> torch.Tensor:
+    if not (-0.5 <= hue_factor <= 0.5):
+        raise ValueError(f"hue_factor ({hue_factor}) is not in [-0.5, 0.5].")
+
+    if not (isinstance(image, torch.Tensor)):
+        raise TypeError("Input img should be Tensor image")
+
+    c = get_num_channels_image_tensor(image)
+
+    if c not in [1, 3]:
+        raise TypeError(f"Input image tensor permitted channel values are {[1, 3]}, but found {c}")
+
+    if c == 1:  # Match PIL behaviour
+        return image
+
+    if image.numel() == 0:
+        # exit earlier on empty images
+        return image
+
+    orig_dtype = image.dtype
+    if image.dtype == torch.uint8:
+        image = image / 255.0
+
+    image = _rgb_to_hsv(image)
+    h, s, v = image.unbind(dim=-3)
+    h.add_(hue_factor).remainder_(1.0)
+    image = torch.stack((h, s, v), dim=-3)
+    image_hue_adj = _hsv_to_rgb(image)
+
+    if orig_dtype == torch.uint8:
+        image_hue_adj = image_hue_adj.mul_(255.0).to(dtype=orig_dtype)
+
+    return image_hue_adj
+
+
 adjust_hue_image_pil = _FP.adjust_hue