Tablet plate name identification method, device and system and tablet plate information acquisition method, device and system

Abstract

The invention discloses a tablet name identification method, which comprises the following steps: carrying out scene text detection on a picture of a tablet to be identified according to a pre-trainedEAST algorithm to obtain all text areas of the tablet to be identified; sorting all the text areas from large to small according to the areas, and determining a plurality of text areas in the front sequence as candidate target areas; performing feature extraction on the candidate target area to obtain image features of the candidate target area; inputting the image features of the candidate target area into a pre-trained long-short-term memory network to obtain a prediction result; converting the prediction result into a text to obtain a predicted drug name; and matching the predicted drug names with drug names in a drug database to obtain the drug names of the to-be-identified tablet plates. The tablet plate medicine name identification method provided by the invention can improve the accuracy of tablet plate medicine name identification. Meanwhile, the invention further provides a tablet plate name recognition device and system and a tablet plate information acquisition method, device and system.

Description

technical field [0001] The invention relates to the field of medical technology, in particular to a method, device and system for identifying drug names on tablet boards and acquiring information on tablet boards. Background technique [0002] Tablet board, that is, the packaging board for pills. Information such as how to use and shelf life of medicines is a serious matter about safety and health. Nowadays, there are many medicines, and the chaotic management of pills has brought inconvenience to people's lives. For some uncommon tablet boards with missing instructions, it is difficult to know a series of information such as the condition of the medicine and precautions for use, so it is easy to cause people to take the medicine by mistake. The existing technology mainly uses traditional OCR (Optical Character Recognition, Optical Character Recognition) technology to identify the drug name of the tablet board. However, for the tablet board with complex text background, tr...

AI Technical Summary

Problems solved by technology

Nowadays, there are many medicines, and the chaotic management of pills brings inconvenience to people's life

For some uncommon tablet boards with missing instructions, it is difficult to know a series of information such as the condition of the medicine and precautions for use, so it is easy to cause people to take the medicine by mistake

The existing technology mainly uses traditional OCR (Optical Character Recognition, Optical Character Recognition) technology to identify the drug name of the tablet board. However, for the tablet board with complex text background, traditional OCR cannot obtain information such as the coordinates of the text, and cannot locate the drug name, resulting in There is too much noise interference and the recognition error rate is high. When the traditional OCR technology cannot recognize the drug name, it is necessary to manually enter the drug name to query the drug information. This is not only cumbersome to operate, but also not conducive to popularization to the elderly.

Images