Xiule Fan

Title: Innovations of Xiule Fan in Stereo Matching Technology

Introduction

Xiule Fan is a notable inventor based in Waterloo, Canada. He has made significant contributions to the field of stereo matching technology, particularly in indoor applications. His innovative approach utilizes deep learning techniques to enhance the accuracy of disparity estimation.

Latest Patents

Xiule Fan holds a patent for a "System and method of unsupervised stereo matching with surface normal assistance for indoor applications." This invention presents a system that employs a deep neural network featuring a feature extraction module, a normal branch, and a disparity branch. The extraction module and the normal branch are initially trained in a supervised manner for surface normal prediction. The predicted surface normal is then integrated into the disparity branch, which is trained later in an unsupervised manner for disparity estimation. This unsupervised learning approach significantly reduces the reliance on extensive ground truth data, which is often challenging to obtain. Experimental results demonstrate that this method can accurately predict surface normals in textureless regions, leading to improved stereo matching quality in indoor scenarios.

Career Highlights

Xiule Fan is currently employed at Avidbots Corp, where he continues to develop innovative solutions in robotics and artificial intelligence. His work focuses on enhancing the capabilities of autonomous systems through advanced algorithms and machine learning techniques.

Collaborations

Xiule collaborates with talented individuals such as Soo Jeon and Baris Fidan, contributing to a dynamic and innovative work environment.

Conclusion

Xiule Fan's contributions to stereo matching technology exemplify the impact of innovative thinking in solving complex problems. His work not only advances the field but also enhances the functionality of indoor applications in robotics.