Edison, NJ, United States of America

Ligong Han

USPTO Granted Patents = 4 

 

Average Co-Inventor Count = 3.5

ph-index = 1

Forward Citations = 1(Granted Patents)


Company Filing History:


Years Active: 2025

Loading Chart...
Loading Chart...
4 patents (USPTO):Explore Patents

Title: Innovations of Ligong Han

Introduction

Ligong Han is an accomplished inventor based in Edison, NJ (US). He has made significant contributions to the field of video synthesis and representation learning. With a total of 4 patents, his work focuses on enhancing video generation frameworks and improving the quality of video representations.

Latest Patents

One of Ligong Han's latest patents is titled "Video synthesis via multimodal conditioning." This patent presents a multimodal video generation framework (MMVID) that utilizes text and images as input, either jointly or separately. The framework employs quantized representations of videos and a bidirectional transformer to predict discrete video representations. It incorporates a new video token trained with self-learning and an improved mask-prediction algorithm to enhance video quality and consistency. Additionally, text augmentation is utilized to improve the robustness of textual representation and the diversity of generated videos. The MMVID framework also integrates various visual modalities, such as segmentation masks, drawings, and partially occluded images, while extracting visual information based on textual prompts.

Another significant patent is "Learning orthogonal factorization in GAN latent space." This method focuses on learning disentangled representations of videos. It involves feeding each frame of video data into an encoder to produce a sequence of visual features. These features are then processed through a deep convolutional network to obtain posteriors of dynamic and static latent variables. The method samples representations from these posteriors and concatenates them for reconstruction. To facilitate disentangled sequential representation learning, orthogonal factorization in generative adversarial network (GAN) latent space is leveraged to pre-train a generator as a decoder.

Career Highlights

Ligong Han has worked with notable companies, including NEC Corporation and Snap Inc. His experience in these organizations has contributed to his expertise in video technology and innovation.

Collaborations

Throughout his career, Ligong has collaborated with talented individuals such as Renqiang Min and Hans Peter Graf. These collaborations have further enriched his work and expanded the impact of his inventions.

Conclusion

Ligong Han is a prominent inventor whose work in video synthesis and representation learning has led to significant advancements in the field. His innovative patents and collaborations highlight his commitment to pushing the boundaries of technology.

This text is generated by artificial intelligence and may not be accurate.
Please report any incorrect information to support@idiyas.com
Loading…