Recently, the 2019 International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI) was held in Shenzhen. Participating teams from Xiamen University (XMU) made remarkable results. One team guided and led by Professor Ji Rongrong from School of Informatics, Xiamen University won four firsts among seven indicators in the Automatic Generation of Cardiovascular Diagnostic Report Challenge (AutoGen-CDR19 Challenge). Another team guided and led by Associate Professor Wang Liansheng won two firsts in the Robust Medical Instrument Segmentation (ROBUST-MIS) of the Endoscopic Vision Challenge.
AutoGen-CDR19 Challenge was initiated by Shenzhen Peng Cheng Laboratory, Tsinghua University and Wuhan Asia Heart Hospital. For many years, ventricular septal defect (VSD) and atrial septal defect (ASD) have become huge threats to human health. Traditional echocardiography technique is widely applied and used in hospital diagnosis. However, for the lack of professionals, and due to the excessive workload and insufficient experience, the reports generated by ultrasound doctors are often rough. Therefore, an intelligent computer-aided report generation system is urgently needed. This challenge aims at evaluating cutting-edge techniques and revealing exciting potentials for medical image captioning, and more particularly, the automatic generation of cardiac-cerebral vascular diagnostic report. The challenge sets seven evaluation indicators (BELU-1, BELU-2, BELU-3, BELU-4, CIDER, METEOR, ROUGE), and requires participating teams to design an intelligent model which can automatically generate the diagnostic report.
The Endoscopic Vision Challenge is held by Heidelberg University and MICCAI. The instrument segmentation in the robotic surgery is an important research problem in the field of robot assisted medicine, and one of the major challenges is to correctly detect the position of the instrument in the surgery for tracking and attitude estimation. Participating teams may enter competitions related to three different tasks, namely binary segmentation, multiple instance detection and multiple instance segmentation. The goal of this challenge is the benchmarking of the medical instrument detection and segmentation algorithms, with a specific emphasis on robustness and generalization capabilities of the methods.
MICCAI is hosted by Medical Image Computing and Computer Assisted Intervention Society, and is a top academic conference across two areas, medical image computing (MIC) and computer-assisted intervention (CAI). Attracting 134 research teams from all over the world, this conference is regarded as greatly influential and authoritative. This year, more than 2,500 participants attended the conference, marking the largest number in history.
