Low-Power CMOS Image Sensor Technologies for Medical Imaging Applications.

The goal of this PhD project is to develop a camera sensor technology to enable software based automatic diagnosis of pathologies in the gastrointestinal (GI) tract. There are many R&D challenges. Present camera image sensors used to take the experimental GI tract videos for study have not been designed for this purpose. Our existing cameras & sensors are large compared to present-day endoscopy cameras; the sensor resolution is low & needs to be increased & the general power consumption is high & needs to be reduced. Such a sensor will need to be tiny compared to what is already on the market & combining this with the high functionality demands will be technologically challenging to achieve. A related challenge is that we presently do not have a camera & sensor which can be introduced into the human body so we will need to create an artificial lab environment in which we can test the sensors' image quality. We need to identify new sensor technologies which will bring out the required amount of detail in the video stream such as: light spectrum, pixel sensitivity, resolution, frame rate & the signal-to-noise ratio. The challenge is to produce an image sensor with the named qualities so that the algorithms can accurately highlight pathologies, giving the least amount of false alarms. If the false alarms are too many, then the resulting product will not be practical to use.

This project is about creating the specifications and documentation for the design of an image sensor which has a high enough image quality to enable the possibility of writing software which will analyse video images, or streams, and in doing so, highlight any abnormal areas in the image(s). The idea is that the software will automatically diagnose pathologies in the gastrointestinal (GI)tract. There are many R&D challenges. Present camera image sensors used to take the experimental GI tract videos for study have not been designed for this purpose. Our existing cameras & sensors are large compared to present-day endoscopy cameras; the sensor resolution is low & needs to be increased & the general power consumption is high & needs to be reduced. Such a sensor will need to be tiny compared to what is already on the market & combining this with the high functionality demands will be technologically challenging to achieve. A related challenge is that we presently do not have a camera & sensor which can be introduced into the human body so we will need to create an artificial lab environment in which we can test the sensors' image quality. We need to identify new sensor technologies which will bring out the required amount of detail in the video stream such as: light spectrum, pixel sensitivity, resolution, frame rate & the signal-to-noise ratio. The challenge is to produce an image sensor with the named qualities so that the algorithms can accurately highlight pathologies, giving the least amount of false alarms. If the false alarms are too many, then our projected pill camera will not be practical to use.