Adaptive and quality-aware storage of JPEG files in the web environment

Abstract

The paper presents a concept, implementation and results of experiments conducted on an online image archiving system which performs JPEG/JFIF compression of user-submitted files. The compression level is chosen automatically based on the quality loss assessment. During development a comparative survey of ten image quality measures, including Mean Squared Error (MSE), Image Fidelity (IF) and Structural Similarity Index (SSIM), has been carried out in order to evaluate their performance in relation to colour image JFIF compression. Furthermore, an Edge Intensity Measure (EIM) has been proposed as a solution to identify images with low edge intensity, whose quality loss is almost universally assessed incorrectly. Finally, an algorithm for determining the most suitable level of compression of a given image has been designed and implemented, incorporating the three image quality measures and the EIM. The proposed algorithm stores the typical JFIF/JPEG files with much higher quality than most of the popular web-based systems, yet without high computing overhead and high image quality deterioration.