Automating my 35mm film scanning pipeline
A developer built an automated Python and local AI pipeline to invert, color-correct, and remove dust from 35mm film scans, reducing a four-hour manual task.
summary is enough The post provides an interesting overview of applying local vision models and numpy to film processing, but the core mechanics fit neatly in a summary.
summary
The author engineered a custom software pipeline to automate the tedious post-processing of 35mm film scans. Utilizing a NumPy script for color inversion, local vision models for metadata tagging, and LaMa inpainting for automated dust removal, the system drastically cuts down manual labor. The finished photos are then published to a custom 3D web gallery built with Three.js. This project bridges traditional analog photography with modern local AI workflows.
key points
- A NumPy routine measures color density using the blank leading slot of a film roll to ensure consistent color inversion across all 36 frames.
- Local Qwen3-VL models running via Ollama automatically tag and index photos within five seconds of scanning.
- An inpainting model detects and removes dust specks while using context awareness to avoid false positives like sun glints.
who should read it
Film photographers and developers interested in applying local AI and computer vision to analog workflows.
discussion positive
consensus
Commenters appreciate the technical ingenuity and automation of a tedious workflow, though opinions are split on whether automating film processing defeats the purpose of shooting analog.
pushback
Traditionalists argue that the deliberate, hands-on labor of developing and scanning film is core to the medium's value, questioning why someone wouldn't just shoot digital instead.
You should check out the negpy project. It in my view has the best inversion process. It claims to emulate the whole analog print chain in the inversion process. They have a fold works cited.
It also automates most of the process. It works with SANE, some other legacy scanners, and camera scanning. It can even do trichrome scans automatically.
It introduces negpy, an existing open-source project that automates film inversion by emulating the analog print chain.
Generated by an LLM from the article and thread. Check the sources before relying on it.