Research On The Lossless Compression Method Of Image Sequences In Cloud Computing

Cloud rendering is the key technology of the next generation three-dimensional animation. In the film production area, rendering special effects or making a3D scene based on computer graphic will generate a massive of image data which will cost most of the network band and storage space. These image sequences will be used to in the next step, so loss compression method is not proper. For the purpose of lossless data compression, we analyze the encoding method of dictionary techniques and then we propose the method of reducing the complexity of the partial data to get more compression, which is based on the rendering generated image characteristics. Current video coding scheme can achieve more compression density, while it is difficult to guarantee absolutely lossless. Besides, the calculate procedure is complex. Image compression method ignore the redundant information of inter frames, which is also not appropriate.To make use of the inter frame redundancy, this paper presents a data block reordering scheme base on LZ77compression method, Which can achieve more compression without no extra expense. What’s more, based on the big correlation between adjacent image frames, a predictive coding model is applied to further reduce data redundancy. Relay on this mind, we present a compression file storage scheme called LZT which is suitable for rendering image sequences.Image hash has the ability to finding similar pixel matrix quickly, which can be used to further improve the compression effect.in order to solve the time consuming problem when processing large-scale of image data, we presents a distributed encoding scheme based on the cloud computing infrastructure.The main work and results are as follows:(1) To the issue of low compression efficiency to the cloud rendering generated image sequences, this article analyses the characteristics of data distribution and LZ-based compression methods in the point of information theory. We put forward a method which increasing the redundancy information of local data to improve the performance. The method makes full use of the overall image sequence redundant information and is suitable for distributed coding. The results show that the compression has a great improvement. (2) To further make use of the inter frame and external frame redundancy, we bring into an prediction coding model which gets improvement from the differential coding. Research the dictionary-based lossless compression method in PNG APNG and MNG format. We put forward a new compression scheme called LZT which is suitable for storage rendering images.(3) Introducing an image hashing mechanism to further improve efficiency. Images can quickly be retrieved through the hash value, and can measure the degree of similarity of the image block. Therefore, the compression performance can also be improved though this method which can get rapid and precisely aggregation.(4) Build a cloud-based distributed lossless coding system to solve computationally intensive tasks with limited processing power in a single computer. The system inherits the advantages of cloud computing platform including robustness, scalability and so on. What’s more the system can make fully use of the data structure, which is still able to maintain a high compression ratio, and parallel computing can significantly reduce the compression time.At the same time, the similar demands can also use this framework such as distributed video transcoding.