I had some pictures with varying resolutions (because of cropping/orientation) that I needed to get to the same aspect ratio. Cropping it in would work, but I didn't want to lose resolution. Instead I wanted to add white borders to pad images. Most existing tools (browser based) required the images to be uploaded and then processed on their server (heavy freemium shilling of course). The one local tool I found could not do it batch.
So I built bother over the course of mostly a weekend. I first used Sharp with Node and had great results, but I couldn't get sharp working on the browser. I wasn't very happy with other existing libraries, so I figured I could use the canvas directly, and it worked out great.
Although all of it is mostly basic, I did have fun figuring out the math required to get the canvas to scale from a specific origin, so that pinch-to-zoom and scroll wheel zoom felt natural. I have the basic idea diagrammed [1] and plan to write about it.
I wrote a small helper [2] that I might release standalone as a library after expanding the feature set.
I plan to add a feature to support slicing. I have polaroid pictures I want to scan, and phone apps don't work great. They also suck for scanning a large number of pictures. I would like to be able to use a flatbed scanner to scan multiple images at once, and then slice them into their own separate images.
I will be making WebGL implementations of canny edge detection, and probabilistic Hough line transforms, which I anticipate to be a lot of fun. This will probably take more than a weekend though.
I have some more ideas, like maybe using generative outpainting instead of white padding? Would love more ideas for what I could add to bother, ideally mundane batch image processing tasks that can be automated.
[1] https://excalidraw.com/#json=VB-95wXb8mmw-WEIe2pNu,Abn9sV1Py...
[2] https://github.com/d4mr/bother/blob/main/src/lib/bother.ts
0