A Preliminary Research On Wireless Capsule Endoscopy Of Image Compress, Convertor And Self-diagnosis

The emergence of wireless capsule endoscopy is an important milestone in the history of the development of digested discipline. Capsule endoscopy uses a swallow capsule that built with wireless video camera for picture taking to assess the gastrointestinal diseases. After moving the images from the image recorder into the imaging workstation, doctors can analyze the health condition of the entire digestive, and thus make a diagnosis. The whole process takes about eight hours, and produces about 50,000 color images of the gastrointestinal tract. Even the experienced doctors have to spend at least two hours to complete each case. The features of capsule endoscopy are a long examination time, mass image data and long time to diagnosis.According to the clinical use of the capsule endoscopy, we could find several drawbacks with this technique. First, the image resolution and filming frequency is low. Second, we could not capture the image through smartphone or network directly. Third, there are a number of capsule endoscopy images, which is hard to deal with.Aiming at the above drawbacks, the work of this thesis is to study image compression algorithm that is easy to implement in hardware to improve the resolution or reducing the amount of data transfer. In order to use a phone or Wi-Fi network to receive the capsule endoscope image, the dual-mode design capsule endoscope image converter is designed. Images can be uploaded directly to the workstation through Wi-Fi network by using the converter. Based on the above study, we finally put forward the self-diagnosis of capsule endoscopy and carry out preliminary verified by experiments. The following is the content of the work in this thesis.First, analysis of the needs and the feasibility of the capsule endoscope image compression. Due to general image sensor has a YUV output format, we analysis the distribution of the capsule endoscope image in YUV space and study the possibility of its implementation of image compression. So we convert 100 images in different parts of the gastrointestinal tract from RGB into YUV to make a statistical difference between the different component values. Analysis the reasons that YUV space is suitable for capsule endoscopy image compression processing. Then we use golomb-rice encoding a capsule endoscope image in the JPEG lossless prediction mode prediction difference entropy encoding, and give the experimental results. Since the UV component in the capsule endoscope image generally has a small deviations, so considered for sampling process. The sampling images are rebuilt and the peak signal to noise ratio is calculated.Second, proposed a method to receive the capsule endoscope images through smart phones and Wi-Fi network. Capsule endoscopy uses a dedicated radio transmission, so it is difficult to use a smart phone or Wi-Fi network to receive data directly. To use the mobile smart device receives capsule endoscope image data or real-time image transfer images to a workstation, this thesis put forward a dual-mode converter based on Bluetooth and Wi-Fi. The converter use a wireless transceiver module the same as the capsule endoscopy to receive the capsule endoscopy images. Depending on the application mode, the microprocessor controlled Bluetooth or Wi-Fi module to convert the data. After the images are compressed by the image compression algorithm, the images can be easily transferred to the smart mobile device or remote network.Third, explore the possibility of achieving self-diagnosis of capsule endoscopy. We carried out preliminary exploration of the self-diagnostic mode for capsule endoscopy after analyzing the deficiencies of the existing capsule endoscopy. We can use a mobile phone to receive the capsule endoscope image directly or upload them to the image workstation by using the converter. After the repeated images and too dark images are deleted by the algorithm, volunteers without medical background pick up the suspicious lesions images from the screened images. And then the images are handed over to doctors to make a diagnosis. Self-diagnostic mode is expected to be as an important supplement mode of capsule endoscopy clinical diagnosis if the self-diagnosis launching a large-scale clinical trial, which will overcome the diagnosis bottleneck effectively.