Jingwei Peng

Title: Innovations of Jingwei Peng in Text-to-Video Localization

Introduction

Jingwei Peng is a notable inventor based in Hangzhou, China. He has made significant contributions to the field of video processing and localization technologies. His innovative approach to text-to-video clip localization has garnered attention in the research community.

Latest Patents

Jingwei Peng holds a patent for a "Progressive localization method for text-to-video clip localization." This method involves extracting features from both video and text modes using different feature extraction techniques. It progressively selects varying step sizes to learn the correlation between video and text in multiple stages. The model is trained in an end-to-end manner based on the correlation loss at each stage. Additionally, the method integrates fine and coarse time granularity stages through a condition feature update module and up-sampling connection, allowing the model to adapt to clips of varying lengths.

Career Highlights

Jingwei Peng is affiliated with Zhejiang Gongshang University, where he contributes to advancing research in video localization technologies. His work has implications for various applications, including multimedia content retrieval and automated video editing.

Collaborations

Jingwei has collaborated with notable colleagues such as Xun Wang and Jianfeng Dong, enhancing the research output and innovation in their projects.

Conclusion

Jingwei Peng's innovative methods in text-to-video localization represent a significant advancement in the field. His contributions are paving the way for future developments in multimedia technologies.