Continuing from last time, we will use ComfyUI to generate images. This time, we will introduce real-time image conversion. The
ComfyUI custom node we will use is ” ComfyUI_toyxyz_test_nodes “. Unfortunately, some functions are not available on Ubuntu. (The functions introduced this time can be used.)
ComfyUI_toyxyz_test_nodesIf
you want to change the image with Image To Image, use the Load Image node to load the image saved on your PC. This node allows you to load images from a webcam or screen capture.
Using the iPhone camera instead of a webcam
This time, I used an app called DroidCam.
(For details, see https://droidcam.app/ )
Connect the node.
Select “image/Webcam Capture” in Add Node.
Image conversion execution ①: The Van Gogh-style
model was created using SDXL Refiner, and the text was created using “The style of Vincent van Gogh, vivid brush strokes.”
Performing image conversion ②: 🍏 → 🍎
The model was Flux.1 dev and the text was “red apple”.
Consideration of image conversion
“Image conversion execution ①: Van Gogh style” turned out as expected, so I think it can be said to be a success.
On the other hand, “Image conversion execution ②: 🍏 → 🍎” was a little tricky, and I ended up with a poorly colored 🍎. In fact, I did it several times, but most of the time it turned out to be 🍏.
If you only want to change the color, it seems better to take the Depth and color it as in the previous article.
