SHANGHAI, Jan. 12, 2018 /PRNewswire/ -- At the beginning of this year, Ping An Technology was ranked as a leader in medical imaging analysis, as its Ping An Yingxiang platform set world records for nodule detection and false positive reduction with reduction with average sensitivity rates of 95.1 percent and 96.8 percent respectively in the authoritative LUNA rankings for medical imaging. It is the first time that Ping An Technology has set world records in medical imaging, demonstrating the company's worldwide leadership in the sector.
LUNA (LUng Nodule Analysis) is not only an internationally recognized evaluation system for pulmonary nodule detection but also one that is receiving a lot of attention across the industry and setting the benchmark for all competitors in the field of medical imaging. The input data used to date comes from several medical institutions in the United States and are annotated by specialists and physicians. The evaluation, held as a form of competition, is being carried out by more than 3,600 professional teams from academic and industrial circles worldwide since 2016, including the Chinese University of Hong Kong, Peking University, Zhejiang University, Alibaba, iFLYTEK, Fosun, as well as a host of startup firms. Each competing team is required to analyze the lung CT samples of nearly 1,000 people and locate the nodules. It is a difficult task as most of the nodules diameters are smaller than 3mm (the same width as three human hairs).
Lung cancer is the leading cause of cancer-related deaths worldwide. Screening high risk individuals for lung cancer with low-dose CT scans is now being implemented in the United States and other countries are expected to soon follow suit. In CT lung cancer screening, millions of scans will need to be analyzed, creating an enormous burden for radiologists. As a result, it has become necessary to develop intelligent algorithms that are optimized for lung cancer detection.
Leveraging its advantages in the relevant areas, Ping An Technology's medical imaging team proposed an innovative solution to this challenge which integrates robust control theory with advanced AI algorithms including deep learning and transfer learning, achieving excellent results and world-leading status in pulmonary nodule detection, localization and recognition. The technology is being applied in collaboration with the Shanghai Institute of Respiratory Diseases and several leading Grade III, Class A hospitals.
Ping An Technology's medical imaging team, led by Dr. Xiao Jing, an expert in China's 1,000 Talents Plan and Chief Scientist at Ping An Technology, is comprised of medical and AI specialists from worldwide top universities including Harvard University, Carnegie Mellon University, Princeton University, the University of Florida, UCLA, Tsinghua University, Zhejiang University and Huazhong University of Science and Technology. The team, with strong R&D backing, now provides a wide range of medical AI capabilities to screen for gastric, liver, cervical and thyroid cancer as well as estimate bone age, detect diabetic retinopathy and thorax related diseases. In terms of the application of medical imaging technologies, Ping An Technology is cooperating with the Shanghai, Shenzhen and Chongqing Health and Family Planning Commissions as well as several of their provincial counterparts, on jointly building platforms based on big data and AI technologies in addition to research into related topics.
Ping An Technology, as the core high-tech division of Ping An Group, develops and manages the key platforms and services that support the development and maintenance of the group's insurance, banking, investment and internet businesses. Ping An Technology also acts as Ping An Group's technology incubator, with strong research, development and application abilities in AI, big data and the cloud.
As Ping An Group continues to grow, Ping An Technology has accumulated extensive experience around its three core strengths - "technology, the internet and finance, and the synergies that the company has built between the three, besides having developed more than 20 leading-edge technologies. In addition to the Ping An Yingxiang platform, the company owns many of what is referred to in Chinese as the "black technologies" - highly-advanced new and emerging technologies - including voiceprint recognition, facial recognition, photo-based damage assessment for motor vehicle accidents, receipt and invoice optical character recognition, voice-to-text, text-to-voice, Q&A chatbot, physical robots and livestock recognition, all of which have applications across a wide range of industries and disciplines. Ping An Technology is a customer-centric organization, where exploration, progress, innovation and the unremitting dedication to achieving breakthroughs in technology are part and parcel of its corporate culture. The company attaches great importance to social responsibility and environmental protection, and aims to become a world-leading financial technology company, enabling a development model based on the premise that, from this point forward, technologies will lead the development of finance.