Previous model-based video coding studies have focused on the modelling techniques themselves and have not considered the transportation of an implementable coder over lossy networks. In this article a hybrid switched 3D model-based/H.261 video coder designed for one-to-many distance learning applications over the Internet multicast backbone (MBONE) is employed to study the susceptibility to data loss of model based packet video. Data is considered as; texture, motion and shape, with further sub-division to periodic and aperiodic types. The absence of the various data streams in isolation is investigated to reveal the types of degradation possible and then statistical loss is applied to all streams simultaneously to observe the combined effects using a simple loss model. The ability of the various streams to recover from these errors is investigated and an order of priority for the data types is then identified.

Modern video compression techniques exhibit variability of performance as a function of time. Studies are reported of viewer reactions to this variability, which indicates a sensitivity to particular features. Some interesting conclusions emerge for future work in video coding.

This paper reports on current developments in the area of model-based video coding, a technique which shows promise of achieving very large bit-rate reductions for moving images. After an introduction and historical review, advances are summarized in several areas, among them improved 3D tracking of the human head and of facial expressions, the use of muscle-driven model animation with skin synthesis, techniques for luminance compensation, and switched coders. Bit rates ranging from 64 kb/s down to about 1 kb/s have been obtained using head-and-shoulder video sequences. Problems with model-based methods are identified and future developments in both CBR and VBR transmission discussed.>

The algorithmic architecture and performance of a video coder are described that switches between the model-based and H.261 modes, depending on image content. The switching criterion incorporates measures of both picture quality and bit rate. The model-based mode uses generalized cylindrical models and a 3D matching technique to estimate motion. The coder was tested with four different image sequences in CBR and VBR operational modes: the performance was generally superior, and in some cases significantly superior, to that of a free-running H.261 coder operating on the same sequences. The switched architecture represents a possible route for the integration of model-based coding with H.261 to give improved performance at very low bit rates.>

Measurements of subjective picture impairment as a function of network loading in a simulated ATM network are reported. The simulation indicated that cells tend to be discarded in bursts, the frequency and severity of which can be related to the loading by a threshold model. The effect of the discards on broadcast-style video, coded using a single-layer H.261-type method, was found to be a function of scene content and movement at the instant of occurrence. If the visibility of cell discards is maintained at or below threshold in worst-case scenes, the study indicated that network loadings around 55% for a multiplex of 16 video sources and around 70% for a multiplex of 48 video sources are achievable.

A study is reported of the extraction of two-level cartoons from moving grey-level images for visual communication in the range 4.8-19.2 kbits/s. It is postulated that perceptually significant features of the human face and hands, at which cartoon lines should be drawn in the image, occur wherever surfaces in object space are approximately tangential to the line of sight of the camera or viewer. The implications of this postulate are analyzed for smooth lambertian reflectors seen against a background of the same luminance factor; it is shown that the combined effects of surface gradient and illumination lead to the formation of luminance valleys at the significant features. An experimental comparison of several detectors is described which confirmed the advantage of valley-based detection for deriving economical but recognizable cartoons.