在通訊技術的蓬勃發展之下，多媒體資料的傳輸需求已經逐漸備受重視了，但是為了能夠有效率的把多媒體資料傳送到接收端，而且還要能夠充分的利用通道的頻寬，就必須對多媒體的資料進行壓縮的動作。壓縮雖然可以達到上述的目的，但是卻引來另一個問題，那就是會造成多媒體資料因傳輸時所引起的錯誤，會在接收端會有錯誤連鎖的反應，也就是說錯誤的部分會遠比實際因通道雜訊所引發的錯誤還來得大，如此便會使得接收端的多媒體資料品質造成嚴重的下降。因此在接收端，都會有一個錯誤恢復的機制，負責解決多媒體資料因通道雜訊所引發的錯誤連鎖效應。 本論文將針對視訊資料的錯誤恢復提出一套低複雜度的演算法並且得到適當的視訊品質。其中結合了前人所提出的方法，並做了一些修改，還有配合利用雙向預測的視訊型態編碼資訊，來選擇該適用哪一種演算法。其主要能夠呈現兩種優點，第一在較小的搜尋範圍之下，能擁有較好的視訊品質。第二，其所呈現的複雜度，會比傳統的邊界比對法(Boundary match algorithm)還來得低 在接收端要執行錯誤恢復機制時，所需要佔用視訊記憶體(frame memory)的程度做一個分析，並且加入了資料重複使用的觀念，使得接收端在做錯誤恢復機制時，能夠提昇使訊資料的使用效率。 The transmission of the multimedia data under internet environment have been wildly used in many applications in the recent but the video data is very sensitive to the transmission error caused by packet lost. This induces decoded video data with error propagation and this effect makes the video quality very poor. In this paper we proposed the algorithm that can make the bidirectional frame as the reference frame in the predictive coding mode under error concealment procedure. By using of this algorithm there are two advantages are presented. The first is to achieve a good recovery video quality with a small search range. The second is to achieve low complexity operation under certain degree of recovery video quality. The error concealment module using proposed algorithm is embedded in the decoder for achieving a certain degree of video quality under error-prone environment. It induces huge requirement of usage of video data to process error concealment procedure. In order to improve the efficiency of video data used in the error concealment module. The data reuse architecture is proposed. The basic concept of data reuse is to construct a local memory in the error concealment module for reducing the access times of frame memory such that improving the efficiency of video data used in the error concealment module.
