Early warning system for dispensing Chinese herbal slices based on video object detection
Objective To construct an early warning system for the dispensing of Chinese herbal slices based on video object detection,and to reduce the incidence of confusion events and ensure the safety of medication.Method Designed an artificial intelligence of Chinese herbal slices dispensing early warning system development scheme,divided into video collection,object detection,optical character recognition and early warning of four modules,early warning system could drive the surveillance video real-time detection of pharmacist Chinese herbal slices dispensing content,real-time comparison with the prescription,wrong dispensing information for voice and graphic information early warning.This article was based on the addition and subtraction of xiaochaihu decoction,3 batches of Chinese medicine tablet samples were purchased,1 524 herbal pieces images were photographed and annotated,and a dataset of Chinese herbal slices was produced.Results After testing,the mean average precision(mAP)of the Faster R-CNN model achieved 95.10%.The trained object detection model combined optical character recognition and early warning algorithm to build an early warning system to identify the process of dispensing Chinese herbal slices,and the system could accurately and automatically alarm the wrong behavior of the dispensing.Conclusions The system can carry out real-time and active detection and early warning of the adjustment of Chinese herbal slices,provide new ideas for the intelligent adjustment of Chinese herbal slices,improve the scientific and technological level of Chinese herbal slices,and promote the combination of traditional Chinese medicine and artificial intelligence.
万方数据
维普
