“AI, Deep Learning, Machine Learning, Computer Vision, Algorithm”

Classification & Detection

Structured deep network based on medical knowledge

Integrate human reasoning principles into CS approach can significantly improve the performance towards clinical problems. Bilateral information can model structural similarity of breasts, which is important in mammogram mass detection. We proposed different deep network (for example: contrasted bilateral network, CBN) for automated mammogram mass detection.

cla.png

Detecting Mammogram

Masses

Column (a)-(c) refer to examined image, contralateral image and ground truth image. Column (d)-(e) present corresponding results by Mask-RCNN and CBN.

Apply Fundings

Wrote Proposal for National Funds and Research Grants

Applying for fundings and research grants from the Ministry of Science and Technology.

Project names:

  1. Multi-scene visual impairment training and visual enhancement technology and product development.(No.SQ2019YFC200257)

  2. Development of medical artificial intelligence product detection platform.(No.YS2019YFB140025)

Write Articles & Figures

Creating network pipeline for CVPR2020.

square.jpg

Company Intro

An AI-based medical imaging diagnostics company, formerly known as the AI Lab Project of Peking University, adn was spun out of PU in 2017. Its AI products are currently used in 350+ China hospitals. Deepwise was rated as 2019 Unicorn Startup.( raised $43 million in C round, led by Sunshine Insurance Group, Changfazhan, SRHC, and the Sunland Fund).