| dc.description.abstract | Wireless capsule endoscopy (WCE) is a non-invasive diagnostic method to investigate diseases in the human gastrointestinal (GI) tract. Specifically, the WCE allows for physicians to visually inspect the GI tract of the patient who swallows the capsule. Thereafter, it transmits video wireless from the inside of the human body to the outside, for approximately eight hours. Because of the size and the limited battery of the capsule, the image quality can be quite poor. By reducing the average bitrate per pixel, the energy consumption will be reduced, which will allow higher quality images, higher framerate or resolution.
When applying the YEF colour transformation instead of the YUV, as well as adaptive prediction, the rate can be reduced by circa 0.12-0.35bpp. Additionally, this resulted in an increase in peak signal-to-noise ratio (PSNR) and Structural Similarity Index (SSIM), depending on the input simulation video. On the other hand, these results were very dependent on the source video and might by different in a real implementation. This should be investigated further with simulation videos more similar to the ones captured by the camera sensor on the capsule.
Another possible solution to improve the overall image quality is to encode some regions of the image in higher quality than the rest of the frame, known as region-of-interest (ROI) coding. A very low complexity, energy efficient, ROI coding scheme was proposed for the encoder (the capsule). The method relies on adaptive sampling rate within a frame, to create a region which isn t down-sampled or filtered. Simulations showed that it works as expected in the encoder, but it does have some problems at the decoder. This should be investigated further, as well as to correctly examine the increase in energy consumption in the proposed method. | |
