No abstract

Quality assessment is of major importance when designing and testing an image/video coding technique. Compression performances are usually evaluated by means of rate-distortion curves. However, the PSNR is commonly employed as the distortion measure. We hereby present a full quality assessment benchmark for the LAR (Locally Adaptive Resolution) coder. We conducted a subjective experiment, where nineteen observers were asked to assess the perceptual quality of LAR coded images under normalized viewing conditions. Furthermore, five objective quality assessment metrics were used in order to determine the most suitable metric for the LAR coder. Finally, both JPEG and JPEG200 images were generated and assessed during the subjective experiment in order to define the optimal quality metric which should be used when comparing the codecs' output images quality.

No abstract

Signals from the three types of cone photoreceptors are generally thought to be combined into two opponent-colour components and an achromatic component. Here we have chosen the cardinal directions Ach (achromatic axis), Cr1 (reddish - greenish axis) and Cr2 (bluish - yellowish axis) defined by Krauskopf to build a colour image coding scheme based on features of the human visual system. In order to design the optimal perceptual quantisers, we studied the perception of colour at different spatial frequencies, measuring the effect of a colour masking signal on the perception threshold for achromatic or chromatic sine-wave gratings. Three cases were considered to describe these interactions. First, we measured the masking effect of each colour component by itself. For the two chromatic components, we observed a pedestal effect for low masking contrasts and threshold rises for higher masking contrasts. Second, we studied interactions between the two colour axes. With colour masking signals composed of Cr1 and Cr2 components, thresholds for Cr1 stimuli dropped at low contrasts of the Cr2 component of the masking signal. This effect disappeared for high Cr2 masking contrasts. This was true at all contrasts of the Cr1 component of the masking signal. Thresholds for Cr2 stimuli were not modified by Cr1 masking. Finally, interactions between the chromatic and achromatic components were more complex. We observed a strong modification of the achromatic contrast-sensitivity function when a Cr1 masking sine-wave grating was added. When the achromatic frequency was lower than the colour masking frequency, the threshold rose. However, when the achromatic frequency was higher than the colour masking frequency, there was a significant pedestal effect. The global effect is a shift of the achromatic contrast sensitivity function. With a Cr2 masking signal, no significant modification of the achromatic contrast-sensitivity function was observed.

No abstract

Through the development of medical imaging systems and their integration into a complete information system, the need for advanced joint coding and network services becomes predominant. PACS (Picture Archiving and Communication System) aims to acquire, store and compress, retrieve, present and distribute medical images. These systems have to be accessible via the Internet or wireless channels. Thus protection processes against transmission errors have to be added to get a powerful joint source-channel coding tool. Moreover, these sensitive data require confidentiality and privacy for both archiving and transmission purposes, leading to use cryptography and data embedding solutions. This chapter introduces data integrity protection and developed dedicated tools of content protection and secure bitstream transmission for medical encoded image purposes. In particular, the LAR image coding method is defined together with advanced securization services.

EROS is the largest database in the world of high resolution art pictures. The TSAR project is designed to open it in a secure, efficient and user-friendly way that involves cryptography and watermarking as well as compression and region-level representation abilities. This paper more particularly addresses the two last points. The LAR codec is first presented as a suitable solution for picture encoding with compression ranging from highly lossy to lossless. Then, we detail the concept of self-extracting region representation, which consists of performing a segmentation process at both the coder and decoder from a highly compressed image, and later locally enhancing the image in a region of interest. The overall scheme provides an efficient, consistent solution for advanced data browsing.

No abstract