Ping An Wins Three Championships at ISBI with its AI Image Technologies
(Hong Kong, 17 April 2019) Recently, the International Symposium on Biomedical Imaging (ISBI) released its 2019 international competition results. Ping An Technology, a subsidiary of Ping An Insurance (Group) Company of China, Ltd. (hereafter “Ping An” or the “Group”, HKEX: 2318; SSE: 601318), won three competitions: Automatic Cancer Detection and Classification in Whole-slide Lung Histopathology (ACDC), Endoscopic Artefact Detection (EAD) and Pathologic Myopia Challenge (PALM). The Company also claimed world firsts in six sub-tasks.
The three competitions won by Ping An Technology at this year’s ISBI conference have significant clinical application value in the medical field. The improvement of image quality addressed by EAD can help doctors improve the accuracy of disease diagnosis and significantly boost the efficiency of image reading, which is of great significance for the development of better clinical man-machine assisted diagnosis and treatment tools and early screening tools.
ISBI is a top international academic conference dedicated to biomedical imaging, co-sponsored by IEEE Signal Processing Society (SPS) and IEEE Engineering in Medicine and Biology Society (EMBS). Every year, ISBI holds an international competition for medical imaging technology worldwide, attracting the participation of world-renowned universities and research institutions (https://grand-challenge.org/).
By virtue of the technical accumulation of millions of medical concepts and medical knowledge maps, professional medical image knowledge, and a powerful platform of AI algorithms, Ping An Technology eventually surpassed its opponents to be crowned the international competition champion.
In the Automatic Cancer Detection and Classification in Whole-slide Lung Histopathology (ACDC) competition, Ping An Technology won the championship with a total score of 0.8373, beating 391 participants.
The competition content of Endoscopic Artefact Detection (EAD) involved the AI automatic detection of endoscopic artefacts. Ping An Technology won with a total score of 0.3610, and also won three sub-tasks of Multi-class artefact detection, Region segmentation and Detection generalisation, beating 478 participants.
The Pathologic Myopia Challenge (PALM) focuses on the detection of AI algorithms associated with the Pathological Myopia in fundus photos. Ping An Technology won the overall championship with the absolute advantage of three core indicators, and won the two sub-tasks of optic disc/cup segmentation and lesion segmentation, beating 345 participants.
Xiao Jing, chief scientist of Ping An Group, said that lung cancer is the most common malignant cancer in the world, and the pathological segmentation of lung cancer has high clinic correlation. The automatic segmentation of lung cancer regions from pathological sections by an AI auxiliary system is helpful in reducing the workload of pathologists and the subjective bias in diagnosis, which is of great significance to the auxiliary diagnosis and analysis of lung cancer.
Xie Guotong, Chief Medical Scientist of Ping An Group
Xie Guotong, chief medical scientist of Ping An Group, said that endoscopy is widely used in early pathological examinations of stomach, colon and other tissues in clinical practice, and endoscopic artifact detection is of great significance for image quality control and clinical diagnosis. Pathological myopia is a retinal degenerative disease, and the incidence rate is increasing year by year in high myopia populations. Early diagnosis and regular follow-ups are very important.
In recent years, Ping An Group has continued to promote the “financial + ecological” and “financial + technology” strategies, relying on leading AI technology to build a powerful repository of medical knowledge and a medical health ecosystem. Based on its technological advantages, Ping An Technology has integrated a complete process solution matrix for end-to-end diseases in the fields of intelligent medical quality control, intelligent disease prediction, intelligent triage and guidance, intelligent medical image screening, intelligent auxiliary diagnosis and treatment. Starting from the pre-diagnosis, diagnosis and treatment of three medical links, Ping An Technology helps solve the medical pain points faced by patients, doctors and users of medical institutions. At present, Ping An Group's medical AI model covers over 1000 common diseases, and about 800 million potential patients benefit from it.