Perceptually optimized low bit-rate image encoding

In this paper we describe a system for high quality encoding of a given image set to a pre-determined, target average Bit-Per-Pixel (BPP). The proposed system uses our proprietary, patent protected, perceptual quality measure to determine the optimal allocation of bits among the images in the image set, and encodes each image using the HEVC/H.265 video encoder with a per image optimal encoding configuration and optional preand post-process. We employ learning methodologies both within the quality measure, and to ascertain optimal per image encoding configurations.