Golden Imaging ( Hong Kong ) Co., Limited (ITF-CS)
GT Medical Systems Limited
With the development of 5G and IoT, edge computing has become a technological development trend. Edge computing can bring many advantages, such as it can provide extremely fast response speed, reduce network load, and ensure the privacy of user data. These advantages make it particularly important in the field of clinical imaging applications, especially for the diagnosis of gastrointestinal endoscopy and ultrasound, doctors need to make judgments and diagnoses during examinations, real-time AI assistance will provide doctors with substantial help. Meanwhile, for doctors with less clinical experience, real-time AI assistance can indirectly enhance their capabilities. Our solution will focus on realizing high efficiency AI on portable edge devices, connecting them to existing medical image equipment for empowerment, so that to achieve multi-edge computing platforms and multi-application scenarios coverage. In terms of technical implementation, on the one hand we will design the dedicated feature-aware deep neural network (DNN) to improve the accuracy of AI-assistant detection; on the other hand, we will use architecture aware optimization and hardware aware optimization for portable edge computing platform to further improve processing speed while maintaining similar accuracy. In this project, we will focus on the application scenario of gastrointestinal (GI) endoscopy, empower AI on the existing endoscopy system, and provide multifunctional real-time auxiliary analysis.