This project uses AIoT technology to face the occurrence issue of incorrect prescription drugs of the drug dispensing phase in hospital dispensing Room. The proposed learning-object-detection based deep technology AI mechanism can automatically identify and verify the correctness of the drugs in the drug bag, thereby significantly improving the accuracy of drug dispensing in Hospital.
This work cooperates with two medical centers to obtain 1,500 prescription drugs, build more than 750,000 image data sets, and use hierarchical multi-model AI drug recognition engine architecture to improve the scalability of the system. The current drug identification accuracy (i.e., Recall) is more than 99％, and response time is less than 0.5 seconds. Compared with the time of 3 seconds required for the current human-visual review by pharmacists, this work can greatly improve the accuracy and efficiency of drug review.
Currently, drug incidents rank first in the number of hospital medical adverse events. This work is in cooperation with two medical centers, and related technology transfer has been completed and commercial development is also underway. In addition to the future application of the work in the hospital＇s dispensing room, the established drug feature image object dataset can also be used as the development of other drug identification applications, which will have a hug impact on the medical industry, medical materials industry, and even the general health industry.

線上展網址：
https://tievirtual.twtm.com.tw/iframe/b5db1150-f9c8-4bf3-b718-13bb1b3aea6a?group=23bfb1fa-dd5b-4836-81a1-4a1809b1bae5&lang=en