Developing Artificial Intelligence Tools for the Next Generation of Medical Endoscopes

The main aim of this research is to explore Artificial Intelligence (AI) techniques to be integrated with the next generation medical endoscope. Consequently, these AI models may be extended to inform the design of the nanostructured optical fibres and aid in the decision making of clinically relevant prognostic schemes. The key objectives of the research are: 1. To develop a novel deep learning network with real-time capabilities to accelerate image reconstruction of optical fibre. The developed model must be scalable and can generalised well in different conditions e.g. temperature change, fibre deformation, etc. 2. Expand image modalities of the endoscope using deep learning techniques to implement a wide range of advanced microscopy in a cost-effective way. 3. To perform extreme data augmentation techniques or implement GAN to increase the diversity and amount of the training dataset for the project. 4. To perform real-time detection and segmentation of abnormalities in endoscopy using deep learning techniques.