Region based image compression using recursive triangular partitioning with a blending model

The proposed novel lossy image compression approach represents an image as segments comprised of variable-sized right-angle triangles. Squares are converted to 4 right-angle triangles, and these triangles are recursively subdivided into smaller triangles until a PSNR quality criteria is met. A novel and economic blending model (similar to Bezier polynomials) is applied to represent the intensities at interior points of the right-angle triangles. Compression and visual quality results compare favorably against a wavelet codec using seven bands.