• Intelligent Precision Medicine Aid System

  • 2019 -11 -07
Pathological whole slide images generally are extremely large. The average size of one image is about 10 to 20 GB. This makes it time-consuming to provide diagnosis and difficult to quantify the analysis results. Traditional automatic digital analysis was used to solve the problems, but the performance was prone to the high variance in tissue appearance. Thus, we build up an Intelligent Precision Medicine Aid System (Alovas), which combines traditional image processing and statistical analysis with artificial intelligence techniques. Alovas can efficiently analyze huge pathological images to speed up doctors’ diagnosis process. Nuclei and lymphocyte detection and counting algorithms as well as Hepatitis and Hepatocellular Carcinoma analysis algorithms are developed and integrated on Alovas. Pathologists (doctors) can browse, compare and label on multiple high-resolution pathological images on this cross-platform (PC and Tablet) software. Also, they can apply our developed automatic analysis to assist their diagnosis. Therefore, Alovas is a powerful computer-aid system for pathological image analysis and diagnosis.
Alovas provides analysis tools including:
(1) Hepatitis analysis. This tool follows the criteria used by professional pathologists, which includes the severity of liver fibrosis and distribution of lymphocytes, and provides Hepatitis scoring and grading for liver pathological images. The accuracy is higher than 90 percent.
(2) Hepatocellular Carcinoma analysis. This tool is able to detect tumor area and conduct further analysis for the area. It provides quantitative data for cell size, shape, numbers and distribution. Furthermore, it uses an algorithm which combines image processing and machine learning techniques to provide Hepatocellular Carcinoma staging.