feat(api): histogram equalization and blending for highres tiling

api/onnx_web/chain/utils.py

@@ -2,7 +2,9 @@
 from math import ceil
 from typing import List, Protocol, Tuple
 
+import numpy as np
 from PIL import Image
+from skimage.exposure import match_histograms
 
 from ..params import TileOrder
 
@@ -38,20 +40,23 @@ def process_tile_grid(
     tile: int,
     scale: int,
     filters: List[TileCallback],
+    overlap: float = 0.5,
     **kwargs,
 ) -> Image.Image:
     width, height = source.size
-    image = Image.new(source.mode, (width * scale, height * scale))
 
-    tiles_x = ceil(width / tile)
-    tiles_y = ceil(height / tile)
+    adj_tile = tile * overlap
+    tiles_x = ceil(width / adj_tile)
+    tiles_y = ceil(height / adj_tile)
     total = tiles_x * tiles_y
 
+    tiles: List[Tuple[int, int, Image.Image]] = []
+
     for y in range(tiles_y):
         for x in range(tiles_x):
             idx = (y * tiles_x) + x
-            left = x * tile
-            top = y * tile
+            left = x * adj_tile
+            top = y * adj_tile
             logger.debug("processing tile %s of %s, %s.%s", idx + 1, total, y, x)
 
             tile_image = source.crop((left, top, left + tile, top + tile))
@@ -60,9 +65,24 @@ def process_tile_grid(
             for filter in filters:
                 tile_image = filter(tile_image, (left, top, tile))
 
-            image.paste(tile_image, (left * scale, top * scale))
+            tiles.append((left, top, tile_image))
 
-    return image
+    scaled_size = (width * scale, height * scale, 3)
+    count = np.zeros_like(scaled_size)
+    value = np.zeros_like(scaled_size)
+    ref = tiles[0][2]
+
+    for left, top, tile_image in tiles:
+        equalized = match_histograms(tile_image, ref, channel_axis=-1)
+        value[
+            top * scale : (top + tile) * scale, left * scale : (left + tile) * scale, :
+        ] += np.array(equalized)
+        count[
+            top * scale : (top + tile) * scale, left * scale : (left + tile) * scale, :
+        ] += 1
+
+    pixels = np.where(count > 0, value / count, value)
+    return Image.fromarray(pixels)
 
 
 def process_tile_spiral(