clipseg VS txt2mask

Following the comments to this old post, I tried to use in-painting with manual mask selection. I didn't get beautiful results but I'm sure with some tweaking, I could make it better. The main problem I had was having to manually select the area where I wanted to place the logo and trying to resize my logo mask to the fit the segment. I tried some automatic segmentation tools (Clipseg and Segment Anything). I couldn't tell the segmentation models to find a good area to for logo placement (i.e. some small flat surface). Given the complexity of what I was dealing with, I think there could be a better way (XY problem).

The addon utilizes clipseg for region masking, which was trained on "an extended version of the PhraseCut dataset" (many thousands of images.)

You'll also need to make sure to update your clipseg repo. The script won't do this for you. Namely you just need to update this models/clipseg.py file to ensure your clipseg has support for the new model.

For localisation at search time I ended up using OWL-ViT. This worked really well. I did not try Detic or CLIPseg but would be interested to hear if anyone else has tried these?

clipseg is an image segmentation method used to find a mask for an image from a prompt. I implemented it as an executor for dalle-flow and added it to my bot yasd-discord-bot.

Automated Replacement (txt2mask) by clipseg

Honestly I'd suggest just downloading the script instead, it's much easier to use and gives you boxes for the prompts rather than having to use all this stuff --> [txt2mask background[/txt2mask] It's up to you of course, but for me the extension conflicts with my favorite extension Dynamic Prompts anyway, so it had to go :( The stand-alone script still works mint tho haha.

Can pix2pix be used with the txt2mask extension to easily isolate the dancer?

That looks similar to Unprompted, which is the successor to txt2mask which is probably much easier to use than the others (at least after a quick glance at the documentation).

automatic1111's webui now has a txt2mask script for inpainting. see here. Works great.

Separating colours into masks then rendering one at a time would just take a plugin, not a model update. It would be like this plugin only a lot easier as the objects are coming pre-masked by the colour instead of having to run recognition on them from the description https://github.com/ThereforeGames/txt2mask

It's pretty awesome but the developer recently said he's wrapping it into another extension of his and will not be updating the standalone script anymore. I'm sad about that because it works differently in the new extension and I find it a lot less convenient now, altho it's possible it will be changed again so I'm not giving up hope yet lol. Here's a link to the original version if you want to try it https://github.com/ThereforeGames/txt2mask

There's a version of this idea already in the Automatic1111 distro, a script called 'txt2mask' that's on github here: https://github.com/ThereforeGames/txt2mask

txt2mask

Perhaps you could txt2mask (https://github.com/ThereforeGames/txt2mask) to automate - i.e. you just need to have the text "fire hydrant" as your mask.