My Projects

3D Vision (2019 - Present)

Leading a 3D vision team, focusing on 3D model reconstruction and AR technologies. Our core 3D algorithms enabled 2 products and 2 internal businesses. The 2 growing products are LinYunJing (an indoor scene reconstruction system) and ZuBangBang (a foot scanning and analyzing system). The 2 businesses are AR shoe trial and virtual broadcast on TaoBao App, which have attracted many top brands/vendors, reached millions of customers, and increased the turnover rate and sales significantly.

Video analytics (2018 - 2019)

Led a video analytics team, developing fundamental CV technologies (video fingerprint/tag/summarization/event detection, face recognition, ReID, OCR, lane detection, etc.) and building a Multimedia AI platform. The platform enabled a few key internal and external businesses.

Augmented biometrics with AI project (2018)

Augmenting face recognition with network location data and facial expression detection provides a more secure and effortless solution for building access and many other applications. I am leading the effort to build a POC with Docker based face recognition backend engine and a frontend app running on iPad. The system was demonstrated to senior leadership in AT&T ATO Ignite Conference in 2018.

Drone video analytics project (2017)

AT&T is using drones for cell tower inspection, and I am leading the effort to build a video analytics platform for indexing, searching, and browsing the large volume of drone videos. Interactive analysis of cell tower inspection drone videos provides two main functions: summarization – where a short video conveys the salient content, and assisted automation of annotation – which reduces the manual effort required for labeling events in videos. This project won the “Leading the Industry” award in AT&T ATO Kickoff Exhibits in 2017.

DIRECTV video analytics (2016)

Augmented the metadata information extracted from DIRECTV content for both real time and offline value-added services.

Automated customer care service (2015)

Massive business knowledge is learned from millions of AT&T home solution and mobility service logs using ML and NLP technologies, and it is used to deliver automated and effortless customer experience.

More...