Medical image recognition method, medical image recognition device, terminal equipment and medium

Abstract

The method is suitable for the technical field of image recognition. The invention provides a medical image recognition method, a medical image recognition device, terminal equipment and a medium. Theinvention discloses a medical image recognition method. The method comprises the following steps: inputting an acquired slice image accommodating a gland form into a trained multi-scale full convolutional network model; feature sets under different receptive fields are extracted from the slice images through the trained multi-scale full convolutional network model; carrying out feature fusion andnormalization processing on the feature set; a feature value set of each pixel point in a slice image can be determined; more pixel point influence factors are considered for the features collected under different receptive fields; according to the method and the device, the overall information of the slice image described by the feature set is more obvious, and the attribution of the pixel points represented by the feature value set is richer, so that the gland cavity position and the gland edge position can be marked in the slice image according to the feature value set of each pixel point,and the recognition efficiency of the medical image is improved.

Description

technical field [0001] The present application belongs to the technical field of image recognition, and in particular relates to a medical image recognition method, a medical image recognition device, a terminal device, and a computer-readable storage medium. Background technique [0002] As important organizational structures in most organ systems, glands are primarily responsible for the secretion of proteins and carbohydrates. Meanwhile, adenocarcinoma is one of the most common forms of cancer, such as prostate cancer, breast cancer, etc. Correct cancer grading can guide professional doctors to recommend appropriate treatment options. Whether the shape of the gland can be accurately identified is crucial for the grading of cancer. [0003] However, in the existing gland image recognition schemes, it is only suitable for identifying a single gland image without lesions, if two or more glands are close together, or the glands have morphological changes due to lesions. Wh...

AI Technical Summary

Problems solved by technology

[0004] In view of this, the embodiment of the present application provides a medical image recognition method, a medical image recognition device, a terminal device, and a computer-readable storage medium to solve the problem of low recognition efficiency in existing gland image recognition solutions

Images