Jingyu Xin

Title: Innovations of Jingyu Xin in Video Detection Technology

Introduction

Jingyu Xin is an accomplished inventor based in Tucson, AZ, known for his contributions to the field of video detection technology. With a focus on automating the sports video editing and highlight generation process, his work addresses the growing demand for efficient event recognition in rapidly evolving digital content.

Latest Patents

Jingyu Xin holds a patent for "Transformer-based temporal detection in video." This innovative approach is designed to recognize and locate events of interest in sports-related videos. The patent outlines a two-stage paradigm that utilizes multiple action recognition models to extract high-level semantic features, followed by a transformer-based temporal detection module to accurately locate target events. This technology has achieved state-of-the-art performance in both action spotting and replay grounding, showcasing its potential for broader applications beyond sports.

Career Highlights

Jingyu Xin is currently employed at Baidu USA LLC, where he continues to develop cutting-edge technologies in video processing. His expertise in machine learning and video analysis has positioned him as a key player in the advancement of automated video editing solutions.

Collaborations

Throughout his career, Jingyu has collaborated with talented individuals such as Zhiyu Cheng and Le Kang. These partnerships have fostered a creative environment that encourages innovation and the sharing of ideas, further enhancing the quality of their work.

Conclusion

Jingyu Xin