PixelArtDetectorConverter will upscale the image if the input image is smaller than the resize sizes.

Added img2img worfklow

PixelArtDetector.py

@@ -162,6 +162,11 @@ class PixelArtDetectorConverter():
         results = list()
         for image in images:
             pilImage = Image.fromarray(np.clip(255. * image.cpu().numpy(), 0, 255).astype(np.uint8)).convert("RGB")
+            resizeBefore = pilImage.width < resize_w and pilImage.height < resize_h
+
+            # resize if image needs upscale
+            if resizeBefore and resize_w >= 128 and resize_h >= 128:
+                pilImage = pilImage.resize((resize_w, resize_h), resample=Image.Resampling.NEAREST)
             
             # Start timer
             start = round(time.time()*1000)
@@ -175,11 +180,11 @@ class PixelArtDetectorConverter():
                     PILOutput = pixelate(pilImage, grid_pixelate_grid_size, transformPalette(palette, "tuple"))
 
             print(f"### {self.CGREEN}[PixelArtDetectorConverter]{self.CEND} Image converted in {self.CYELLOW}{round(time.time()*1000)-start}{self.CEND} milliseconds")
-            
-            # resize
-            if not isGrid and resize_w >= 128 and resize_h >= 128:
-                PILOutput = PILOutput.resize((resize_w, resize_h), resample=Image.Resampling.NEAREST)
 
+            # resize if image needs downscale
+            if not resizeBefore and not isGrid and resize_w >= 128 and resize_h >= 128:
+                PILOutput = PILOutput.resize((resize_w, resize_h), resample=Image.Resampling.NEAREST)
+                
             # Convert to torch.Tensor            
             PILOutput = np.array(PILOutput).astype(np.float32) / 255.0
             PILOutput = torch.from_numpy(PILOutput)[None,]

README.md

@@ -1,4 +1,4 @@
-# ComfyUI PixelArt Detector v1.1
+# ComfyUI PixelArt Detector v1.2
 Generate, downscale, change palletes and restore pixel art images with SDXL.
 
 ![](./examples/Image_00135_.webp) ![](./examples/Image_00157_.webp) ![](./examples/Image_00162_.webp) ![](./examples/Image_00165_.webp) ![](./examples/Image_00166_.webp)
@@ -9,7 +9,9 @@ Generate, downscale, change palletes and restore pixel art images with SDXL.
 
 Save a picture as Webp (+optional JPEG) file in Comfy + Workflow loading.
 
-**Update**: changed the default workflow.json to use the default "Save Image" node. workflow_webp.json will be using the webp node.
+**Update 1.2**: PixelArtDetectorConverter will upscale the image if the input image is smaller than the resize sizes.
+
+**Update 1.1**: changed the default workflow.json to use the default "Save Image" node. workflow_webp.json will be using the webp node.
 
 > [!IMPORTANT]
 > If you have an older version of the nodes, delete the node and add it again. Location of the nodes: "Image/PixelArt". I've added some example workflow in the workflow.json. The example images might have outdated workflows with older node versions embedded inside.
@@ -62,6 +64,8 @@ Set the resize inputs to 0 to disable upscaling in the "Save" node.
 
 Reduce palettes or completely exchange palettes on your images.
 
+##text2image
+
 **Positive prompt**: pixelart, {your scene}, pixel-art. low-res, blocky, pixel art style, 8-bit graphics, sharp details, less colors, early computer game art
 
 **Negative prompt**: sloppy, messy, blurry, noisy, highly detailed, ultra textured, photo, realistic, high-resolution, photo-realistic
@@ -70,6 +74,26 @@ Reduce palettes or completely exchange palettes on your images.
 ![](./examples/community/PixelArt_00004_.webp) ![](./examples/community/ImageConverted_00006_.webp)
 ![](./examples/community/ImageConverted_00005_.webp)
 
+##image2image
+
+on a pixel drawing converting to a pixel asset for finetuning:
+
+Input:
+![](./examples/pixil-frame-0.png)
+
+Output:
+![](./examples/PixelArt_00065_.webp) 
+![](./examples/PixelArt_00058_.jpg)
+![](./examples/PixelArt_00071_.webp)
+
+Workflow WEBP (drag & drop):
+![](./img2img_webp.png)
+
+Feel free to experiment!
+
+Other image2image examples:
+https://github.com/dimtoneff/ComfyUI-PixelArt-Detector/issues/2
+
 ### Extra info about the "PixelArt Detector (+Save)" Node:
 
 There is a compression slider and a lossy/lossless option for webp. The compression slider is a bit misleading.