Stanford University-See More in Images by Artificial Intelligence

Keywords​:

Deep Learning; image restoration; computer vision;

Desired candidates​: Students with basic programming skills with general knowledge of calculus and

linear algebra. Students of all majors are welcome.

Introduction​:

The goal of this project is to use artificial intelligence tools to discover more information in images with degradations. These degradations appear in different imaging modalities, from the normal camera photography to medical imaging. This project is designed to provide essential backgrounds and tools, and encourage students to develop new ​machine learning technologies to solve this challenging problem. More specifically, the following topics will be covered mainly:

• Natural image restoration​: focus on how to recover the object structure from noisy natural images, for instance, the photos taken from mobile phones (Fig. 1).
• Medical image processing:​ develop approaches to improve the image quality of medical images, potentially having significant impacts on the early diagnosis of diseases (Fig. 2).

Schedule​:

Week 1: foundations of machine learning, especially deep learning, and image restoration basics Main results:​

get ready for the development of image restoration techniques using machine learning tools.

Week 2: natural image restoration Main results:​

Be familiar with the state-of-the-art techniques for the restoration of images taken by mobile phone, and potentially propose new methods.

Week 3: medical image enhancement Main results:​

Apply and develop approaches to significantly reduce the image artifacts from medical setting.

Week 4: conclusion and ending with reports.