leehomyc/Faster-High-Res-Neural-Inpainting

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

1286 stars · 207 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

Research code implementing the CVPR 2017 paper 'High-Resolution Image Inpainting using Multi-Scale Neural Patch Synthesis', written in Lua for the Torch framework. It fills in missing regions of images using a content network plus texture optimization, with pre-trained models and demo scripts included.

Use cases

• fill in missing or masked regions of high-resolution images
• remove unwanted objects from photos using neural inpainting
• reproduce CVPR 2017 multi-scale neural patch synthesis results
• run image inpainting experiments with pre-trained Torch models
• photo editing via content-aware fill

When to choose

• you need to reproduce or build on this specific CVPR 2017 inpainting paper
• you work in Torch/Lua and want a research-grade inpainting implementation
• you want high-resolution inpainting with texture optimization rather than a quick GAN fill

When to avoid

• you need a maintained, production-ready inpainting tool
• your stack is PyTorch or TensorFlow rather than Torch/Lua
• you need fast real-time inpainting (original version took ~30s/image even after 6x speedup)

Facets

library · maturity abandoned

image-processing deep-learning machine-learning computer-vision image-processing deep-learning machine-learning lua image-inpainting torch research-code cvpr-paper neural-patch-synthesis gpu linux

Member repositories

For agents

markdown · JSON · MCP: product_card(name="leehomyc/Faster-High-Res-Neural-Inpainting")