Context Quantization Based On Multi - Mode Optimization Clustering Algorithm

In terms of data compression problem, entropy coding algorithm has been occupying an important position. However, when dealing with high-order entropy coding problem, the model is a cost issue can’t be ignored. Context model originally proposed to improve coding compression performance and reduce code length coding problems. But when dealing with the issue of higher order Context dilution, what it brings a very weak advantage. To solve the model dilution problem, can it be too dispersed clustering statistics, and ultimately through the Context model to quantify the improved compression performance.This paper presents a new clustering algorithm-MOA Context model based clustering algorithm to quantify the problem in order to reduce the Context dilution model, thus improving the overall compression performance. Because Multimode optimization clustering algorithm is a new algorithm, which is based on the theory MOA algorithm. MOA algorithm is an innovative discrete optimization algorithm that is through the entire idea of local and global search to realize careful search, as the same time by means of learning and memory structure of the table, to get more of the outstanding experience gained on the basis of solution. The multimode optimization clustering algorithm used the MOA algorithm and the idea of clustering to achieve, so the algorithm can play a natural advantage of the MOA algorithm and achieve the clustering capabilities.The multimode optimization clustering algorithm achieved context quantification in this article. First of all, a selected image is clustered to obtain the mapping table, the mapping table obtained as the training set; then choose another image as a set of experiments conducted to verify the feasibility of this encoding algorithm, by calculating the number of bytes consumed encoding normal circumstances compare the advantages of this algorithm certified; and finally by the decoding process to restore the image to confirm the feasibility of the algorithm. Experimental part highlighted the impact of the number of clusters on the coding results, a large number of experimental data was found by statistical clusters close to the actual number of33, and proved with respect to unconditional coding, codingefficiency is increased about5times. Multimode optimization clustering algorithms incomputational complexity, time-consuming, stability, accuracy and other aspects ofthe advantages, but also on the context quantification plays to their strengths.