Wisdom tooth identification and classification method based on deep learning network

Abstract

The invention provides a wisdom tooth recognition and classification method based on a deep learning network, and relates to the technical field of wisdom tooth image processing, and the method comprises the steps: S1, obtaining a panoramic X image, constructing an image data set, carrying out the preprocessing, marking wisdom teeth, and obtaining a position label; s2, dividing the processed and labeled data set, and constructing and training a Yolov5 network model; s3, inputting an image into the model, identifying a wisdom tooth region and segmenting the wisdom tooth region to obtain a local wisdom tooth feature map; s4, marking the dental crown surface of the wisdom tooth according to the local wisdom tooth feature map to obtain a marking point set of the dental crown surface; and S5, carrying out linear regression on the marking point set, calculating the direction of the tooth axis, and carrying out classification according to a set threshold value. According to the invention, wisdom tooth recognition is realized through the Yolov5 and Resnet18 network models, the recognition efficiency and accuracy are improved, and the tooth growth direction is obtained while wisdom tooth recognition is carried out.

Description

Technical field [0001] The invention involves the field of wisdom tooth image processing technology, and specially involves a wisdom tooth recognition and classification method based on deep learning networks. Background technique [0002] Wisdom teeth are also known as the third grinding teeth. They often sprout after the age of 16. Because the last germination of teeth in the mouth, most of the sprouting space and abnormal morphology, the reasons for the direction of the eruption cause their obstructions can not eject normally, and the incidence rate can reach 80 %. Due to the third molar obstruction, the incidence rate is 7 % -78.4 % due to the inability to clean the neighboring teeth; Most of the wisdom teeth are mostly removed in time after discovery. [0003] Since the third molar has developed from the age of 11, after the development of the dental root at about 17 years old, it can be determined whether it can sprout normally according to its growth direction and establis...

AI Technical Summary

Problems solved by technology

[0004] At present, the identification of wisdom teeth is generally judged by direct observation of oral panoramic X-rays relying on the experience of doctors, which has considerable limitations for the identification of wisdom teeth; in the prior art, there are also models constructed by neural networks for identification of tooth types. However, for teeth, different types of teeth have different characteristics. The identification of wisdom teeth is different from the identification of other types of teeth. The identification of wisdom teeth is to facilitate subsequent timely extraction, and the extraction of wisdom teeth also needs to identify the growth of wisdom teeth direction, the neural network for tooth recognition in the prior art has a low recognition accuracy for wisdom teeth, and cannot obtain the growth direction of wisdom teeth

Images