Xingbo Guan

Title: Innovations of Xingbo Guan in Image Processing

Introduction

Xingbo Guan is a notable inventor based in Hong Kong, CN. He has made significant contributions to the field of image processing, particularly in enhancing low-light images through innovative neural network techniques. His work is characterized by a deep understanding of convolutional neural networks and their application in real-world scenarios.

Latest Patents

Xingbo Guan holds a patent for a neural network designed for raw low-light image enhancement. This invention involves a neural network that receives mono-color pixels in a Bayer pattern from an image sensor. It interpolates these pixels to generate full-color RGB pixels while also enlightening the image for better detail. The patent describes a Back-Projection (BP) layer added to each contracting layer in a U-net convolution neural network, where feature depth is increased as pixels are downsampled. Additionally, a BP channel shrink layer is incorporated into each expansion layer to upsample and increase pixel resolution. The convolution layers in the BP layer are responsible for lightening and darkening images to generate an error, which is then lightened and added to a weighted lightened input. The neural network learns the optimal weightings to correct for noise and error in the lightening process within these BP layers. Noise is further reduced by repeating a convolution and leaky Rectified Linear Unit (lrelu) in series for the first convolution in the BP layer that performs lightening.

Career Highlights

Xingbo Guan is affiliated with the Hong Kong Applied Science and Technology Research Institute, where he applies his expertise in image processing and neural networks. His innovative approach has led to advancements in how low-light images are processed and enhanced, making significant impacts in various applications.

Collaborations

Xingbo has collaborated with notable colleagues such as Wan-Chi Siu and Yan Huo, contributing to the advancement of technology in their field. Their combined efforts have fostered a collaborative environment that encourages innovation and research.

Conclusion

Xingbo Guan's contributions to image processing through his patented neural network technology exemplify the potential of innovation in enhancing visual data. His work not only addresses practical challenges in low-light image enhancement but also paves the way for future advancements in the field.