Useful to generate a pixellated image for example.

Add a background mask to an SVG icon

Draw it pixel by pixel in this web UI

Utiliser WebP:

• png : 477,8 KiB
• jpg : 103,9 KiB
• webp : 51,0 KiB
• webp : 40,3 KiB (avec flou gaussien appliqué de 0,5px)

Les services en lignes permettent de gagner quelques pourcentages grâces à des presets plus aggressifs.

• png : 171,1 KiB (-61%) via https://tinypng.com/
• jpg : 97,5 KiB ( -6%) via https://tinyjpg.com/
• webp : 49,5 KiB ( -3%) via https://ezgif.com/optiwebp
• webp : 38,8 KiB ( -4%) (avec flou gaussien appliqué de 0,5px)

Generate beautiful mesh gradients

deface is a simple command-line tool for automatic anonymization of faces in videos or photos. It works by first detecting all human faces in each video frame and then applying an anonymization filter (blurring or black boxes) on each detected face region. All audio tracks are discarded as well.

PaddleOCR aims to create multilingual, awesome, leading, and practical OCR tools that help users train better models and apply them into practice.

It has support online. It seems to a a promising project!

Pour vos photos:

• limitez les dimensions, un maximum de 2 000 x 2 000 pixels est largement suffisant;
• utilisez le format JPEG (ou AVIF si vous voulez promouvoir le format du futur);
• réglez la qualité à 75 ou moins (35 dans le cas du format AVIF).

Pour des images non-photoréalistes:

• en général vous pouvez réduire encore plus les dimensions que pour les photos;
• utilisez le format PNG.

Je vais jeter un coup d'oeil au format AVIF :)

Le post décrit ensuite certaines définitions, et comment compresser manuellemenet

Automate screenshots of websites with this tool for the command-line interface.

An introduction (fr): https://korben.info/shot-scraper-capturer-site-web.html

Lychee is a free photo-management tool, which runs on your server or web-space. Installing is a matter of seconds. Upload, manage and share photos like from a native application. Lychee comes with everything you need and all your photos are stored securely.

Draw vectorial images

and the project: https://github.com/methodofaction/Method-Draw

• DALL-E 2:
• craiyon is a mini DALL-E opensource
• text2image: generate an image from text prompt. " Predictions run on Nvidia T4 GPU hardware. Predictions typically complete within 6 minutes. The predict time for this model varies significantly based on the inputs.
• Wombo: generate an image from a maximum of 200 characters and an art style. An image can be provided as "inspiration" too. It supports NFTs o_o

Improve image quality with AI. One mode is for drawing and another for photos.

It is efficient to improve quality lost compression, removing color branding and vectorizing an image.

An image compression library in javascript based on the Squoosh tool

Convert to AVIF

This app doesn’t seem to have any limitations on the number of files or file sizes, but note that the AVIF format is not yet supported in all modern browsers.

Optimizes .mp4, .web and .gif

100% client-side