# PhotoAIBench

> Measured tests of AI photo tools: upscalers, background removers, and restoration.

Written by one author, a Principal ML engineer (10+ years in machine learning, from research to production; Former Staff ML Engineer working on document extraction and OCR; Former Lead ML Platform Engineer for a 5,000-stream computer-vision platform; Contributor to vLLM; creator of Videoflow (1,000+ GitHub stars)). The tools run in the browser and never upload files. Every page has a Markdown version at the same path with .md appended.

## Tools

- [Batch image resizer and compressor: JPEG, WebP, AVIF, metadata removed](https://photoaibench.com/image-resizer-compressor.md): Resize many photos at once by longest side, width, height or percent, re-encode them as JPEG, WebP or AVIF with a quality slider that shows the real file size, and strip EXIF and GPS data. Runs in your browser; nothing is uploaded. Includes measured file sizes and quality for each setting.
- [PSNR and SSIM calculator: compare two images in your browser](https://photoaibench.com/psnr-ssim-calculator.md): Load an original and a processed copy of the same size and get PSNR (RGB and luma), SSIM per Wang et al. 2004, and a heat map of where they differ. Runs in your browser; images are never uploaded. The formulas and the exact settings are on the page.

## Benchmarks and comparisons

- [AI upscaler baselines: bicubic, Lanczos and Real-ESRGAN on 12 public-domain photos](https://photoaibench.com/upscaler-baselines.md): Twelve public-domain and CC0 photos reduced 4x and enlarged again by five free methods, from bicubic interpolation to Real-ESRGAN, scored against the originals with PSNR, SSIM and LPIPS, with side-by-side crops. Not a ranking of products; the yardstick commercial upscalers have to beat.

## Guides

- [How to choose an AI upscaler (and test one on your own photos first)](https://photoaibench.com/how-to-choose-an-ai-upscaler.md): What actually differs between AI image upscalers, from measured tests on public-domain photos. The trade-off between faithful and invented detail, how to test any upscaler on your own photos in ten minutes before paying, what free tiers and terms of service hide, and when free open-source tools are enough.
- [How PhotoAIBench tests AI photo tools](https://photoaibench.com/methodology.md): The rules behind every PhotoAIBench test. How the public-domain and CC0 test photos are chosen and licensed, how an upscaler is measured, what PSNR, SSIM and LPIPS mean in plain words, how commercial tools are tested within their terms, and the background-removal and restoration tests that are planned but not yet run.

## Optional

- [About PhotoAIBench and its author](https://photoaibench.com/about.md): Who runs PhotoAIBench, the computer-vision background behind its tests, how the site is funded, and how to report a wrong license, source or number.
- [Affiliate disclosure](https://photoaibench.com/affiliate-disclosure.md): How PhotoAIBench may earn money from affiliate links, how those links are labeled, the rules that keep commissions out of test results, and the current list of relationships.
- [Privacy](https://photoaibench.com/privacy.md): PhotoAIBench's image tools run in your browser and never upload your photos. The site sets no cookies and has no accounts. Analytics, if switched on, is cookie-free.