Xiaohui Shen

San Jose, CA, United States of America

Xiaohui Shen

This inventor holds 104 USPTO granted patents and 3 published patent applications, primarily in Image Processing (CPC class G06T2207-20084). Top assignees: Adobe, Inc., Nokia Corporation, Nokia Technologies Oy. Active years: 2013-2025.

USPTO Granted Patents = 104 

% Patents Active = 99.0

Average Co-Inventor Count = 4.4

ph-index = 13

Forward Citations = 605(Granted Patents)

Forward Citations (Not Self Cited) = 529(Dec 10, 2025)


Inventors with similar research interests:

🔬 See more inventors

Location History:

  • Evanston, CN (2015)
  • Evanston, IL (US) (2013 - 2019)
  • San Jose, CA (US) (2015 - 2024)

Company Filing History:


Years Active: 2013-2025

Loading Chart...
Areas of Expertise:
Image Enhancement
Neural Network
Object Detection
Image Cropping
Deep Learning
Image Segmentation
Illumination Estimation
Digital Image Completion
Patch Matching
Aesthetics-Guided
Visual Similarity
Boundary-Aware
104 patents (USPTO):Explore Patents

Xiaohui Shen is a well-known inventor in the field of digital image processing and has been contributing significantly to this field from his base in San Jose, California, for many years now. He has filed a total of 100 patents to date, which is a testament to his expertise and dedication.

His latest work includes two patents that are based on innovative techniques related to digital image completion and frame selection using a trained neural network. In one of his latest inventions, Xiaohui Shen has developed a technique for selecting frames based on a trained neural network, specifically a convolutional neural network. The selection process involves using a loss function that is based on the estimated quality difference between two training frames from a frame collection. In addition, the neural network is also trained using facial heatmaps generated from the training frames and quality scores of detected faces.

In his other patent, Xiaohui Shen has developed a technique for digital image completion by learning generation and patch matching jointly. The technique leverages a dual-stage image completer framework that combines a coarse image neural network and an image refinement network to generate a filled digital image with refined imagery. The image refinement network utilizes a patch matching technique, filtering patches generated based on the coarse prediction with information from patches of known pixels.

Xiaohui Shen has worked for various organizations, including Adobe Inc. and Adobe Systems Inc., contributing significantly to the field of digital image processing. He has collaborated with other experts in the field, such as Zhe Lin and Radomir Mech, to develop innovative techniques and technologies that have pushed the boundaries of digital image processing. With his latest patents, Xiaohui Shen continues to demonstrate his unwavering commitment to innovation and his invaluable contribution to the field.

This text is generated by artificial intelligence and may not be accurate.
Please report any incorrect information to [email protected]
Loading…