Texture and image interpolation using Markov models
Technion - code: COM-1071
Image upscaling technologies are in demand due to the influx of high-definition displays and the prevalence of low-resolution content. Our algorithm allows for the restoration of textured images out of low-resolution or low-quality ones. The algorithm analyzes the image statistically to achieve superior results as compared to simple methods such as bicubic and bilinear. This also enables high-ratio compression for textures and introduces an innovative criterion for comparison of image textures, which may be used for different applications.