Image Processing Glossary.
Clear technical definitions for alpha channels, neural segmentation, WebAssembly, ONNX runtimes, and digital image formats.
Alpha Channel
A color channel that specifies pixel opacity in digital images, allowing values from 0 (completely transparent) to 255 (completely opaque).
Image Segmentation
A computer vision process of partitioning a digital image into multiple pixel regions based on semantic category or object boundary.
Image Matting
The process of calculating precise fractional alpha opacity values for fine structural features like hair strands, fur, and glass reflections.
PNG (Portable Network Graphics)
A raster graphics file format that supports lossless data compression and full 8-bit alpha channel transparency.
WebP
A modern image format developed by Google offering superior lossy and lossless compression for web images with alpha transparency.
ONNX (Open Neural Network Exchange)
An open format built to represent machine learning models, enabling model interoperability between PyTorch, TensorFlow, and browser runtimes.
WebAssembly (WASM)
A low-level binary instruction format for web browsers enabling high-performance, near-native execution of code written in C, C++, and Rust.
Transparent Background
An image canvas state where scene backdrop pixels are removed, represented visually by a grey-and-white checkerboard pattern.