No abstract

Asynchronous Transfer Mode (ATM) networks are well-suited for video because they can directly transmit variable bit rate video signals. Fuxther, ATM networks can carry several services such as voice, video, and data using the same hardware and protocols. The Lapped Orthogonal Transform (LOT) has been proposed for video coding applications in ATM networks. Data-overlapping performed by the LOT algorithm can reduce greatly the visible effect of lost ATM cells; information f " the neighbors of lost cells automatically helps replace any missing data. This paper proposes and evaluates further techniques for hiding defects caused by lost blocks of LOT data. The mathematics of LOT calculation suggest several reconstruction strategies. Standard signal processing techniques such as mean-square-emr minimization, time-domain filtering, and LOT-domain coefficient averaging are also investigated and simulated. The paper's conclusion comments on the performance of these recovery methods for moving video sequences as well as for single frames.

No abstract

No abstract

No abstract