Add files via upload

__init__.py

@@ -0,0 +1,12 @@
+
+
+from .sbs import SideBySide
+
+NODE_CLASS_MAPPINGS = {
+    "SBS_by_SamSeen": SideBySide
+}
+
+NODE_DISPLAY_NAME_MAPPINGS = {
+    "SBS_by_SamSeen": "👀 Side By Side"
+}
+ 

requirements.txt

@@ -0,0 +1,3 @@
+numpy
+Pillow
+torch

sbs.py

@@ -0,0 +1,78 @@
+import torch
+from PIL import Image
+import numpy as np
+import tqdm
+
+
+class SideBySide:
+    def __init__(self):
+        self.device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+
+    @classmethod
+    def INPUT_TYPES(cls):
+        return {
+            "required": {
+                "base_image": ("IMAGE",),
+                "depth_map": ("IMAGE",),
+                "depth_scale": ("INT", {"default": 50}),
+            },
+            
+        }
+
+    RETURN_TYPES = ("IMAGE",)
+    FUNCTION = "SideBySide"
+    CATEGORY = "👀 SamSeen"
+
+    def SideBySide(self, base_image, depth_map, depth_scale):
+        """
+        Create a side-by-side (SBS) stereoscopic image from a standard image and a depth map.
+
+        Parameters:
+        - base_image: numpy array representing the base image.
+        - depth_map: numpy array representing the depth map.
+        - depth_scale: integer representing the scaling factor for depth.
+
+        Returns:
+        - sbs_image: numpy array containing the stereoscopic image.
+        """
+        # Convert tensor to numpy array and then to PIL Image
+        image_np = base_image.cpu().numpy().squeeze(0)  # Convert from CxHxW to HxWxC
+        image = Image.fromarray((image_np * 255).astype(np.uint8))  # Convert float [0,1] tensor to uint8 image
+
+        depth_map_np = depth_map.cpu().numpy().squeeze(0)  # Convert from CxHxW to HxWxC
+        depth_map_img = Image.fromarray((depth_map_np * 255).astype(np.uint8))  # Convert float [0,1] tensor to uint8 image
+
+        # Get dimensions and resize depth map to match base image
+        width, height = image.size
+        depth_map_img = depth_map_img.resize((width, height), Image.NEAREST)
+
+        # Create an empty image for the side-by-side result
+        sbs_image = np.zeros((height, width * 2, 3), dtype=np.uint8)
+        depth_scaling = depth_scale / width
+
+        # Fill the right half of the SBS image with the base image
+        for y in range(height):
+            for x in range(width):
+                color = image.getpixel((x, y))
+                sbs_image[y, width + x] = color
+
+
+        # Apply depth map to shift pixels in the left half of the SBS image
+        for y in tqdm.tqdm(range(height)):
+            for x in range(width):
+                depth_value = depth_map_img.getpixel((x, y))[0]
+                pixel_shift = int(depth_value * depth_scaling) + 1
+                if x + pixel_shift < width:
+                    new_coords = (x + 1, y) if x + 1 < width else (x, y)
+                    for i in range(pixel_shift):
+                        sbs_image[y, width + x - i] = image.getpixel(new_coords)
+
+        # Copy the base image to the left half of the SBS image
+        for y in range(height):
+            for x in range(width):
+                color = image.getpixel((x, y))
+                sbs_image[y, x] = color
+
+        # Convert back to tensor if needed
+        sbs_image_tensor = torch.tensor(sbs_image.astype(np.float32) / 255.0).unsqueeze(0).unsqueeze(0)  # Convert back to CxHxW
+        return sbs_image_tensor