Wireless Endoscope System And Endoscope Image Compression

Wireless endoscopes are capable of eliminating the limitations of conventional GI endoscopes. Subminiature camera capsule could be swallowed into the intestine to take photographs of specific and non-specific locations of the gullet,stomach,large and small intestines. This is a great advantage over traditional endoscopes,which can only examine the upper and lower GI tracts. It is very important for health care area and the diagnoses of the alimentary tract illnesses to research the wireless endoscope system and its relative image processing technology. But, because of the quality problems resulted from the storage, transmission and compression algorithm of image, the future usage of wireless endoscopes get great influences.The intrinsic information redundancy and statistic characteristic of the image sampled from wireless endoscopes systems make that of endoscopic images datas' compression become possible and provide foundations for image compression. The visual attribute provides conditions for further compression of sampled image. We can use one of those two conditions mentioned above for lossy and lossless compression, or we can combine those two foundations for efficient compression.The main contents and contributions of this paper are summarized as follows:â‘ The evolution and actuality of similar wireless capsule endoscope systems are thoroughly researched. The system overall scheme and system composition are designed based on these structures of both home and abroad. Image pickup module, wireless transceiver circuit and wireless transceiver and storage system outside of bady are also designed.â‘¡The model, structure, classification and learn rules of the BP net are discussed and deeply researched, BP algorithm and the theory of the data compression is studied and the application in image compression of BP net and its key technologies are further researched too. But in real application, when classical BP ANN is applied to image compression, it contains those follow shortcomes, low convergence rate when net is trained, easily getting into local minimum and weakly global seaching ability. This article puts forward a kind of BP NN based on GA, and it is applied to image compression. Base on the characteristic that bigger probability of finding global solution in the whole variable space of GA and the property of BP algorithm, the new GA-BP presents better converge speed and better robust. The result of image compression experiment shows higher quality of recovery image. This is also one of innovations of this article.â‘¢Based on requirement analysis and feasibility analysis of compression of medical image, this article systemly studys the basic theory knowledge of ANN and Multi-waveletlet, and construct a new kind of image compression system. In this system, BP net and Multi-waveletlet are the two main technologies. Among this system, image processing is in YCbCr space and the pre-filter technology of Multi-waveletlet is improved, so the enegy of image after Multi-waveletlet transform is better concentrated. A new compression method is introduced to the high frequency and low frequency parts of the image after Multi-waveletlet transform, that is that JPEG is used to encode low frequency part of image to improve compression ratio, and GA-BP net is used to compress and encode high frequency part, so better performance is attained. In the whole system, the usage of Multi-waveletlet mix-construction technology is the main part and that make it better than other encoder systems. Among those three experiments of Multi-waveletlet mix-construction, the low frequency part of GHM Multi-waveletlet transform and the high frequency part of CL is used in three kinds of image compression systems (Smooth, streak and neutral), and this mixture not only better realizes the construction of image, but also produces better compression effect, than BP net, Multi-waveletlet or wavelet encoder system alone.Although a non-invasive gastrointestinal inspecting system has been developed by the author, there still exists much research work to be done to make the system suitable for clinical applications. At the latter of this paper, a summary and a prospect for all the research are presented and the future works for the project are introduced.