Automatic detection method and system for lesion area in pathological tissue slice image

Abstract

The invention discloses an automatic detection method and system for a lesion area in a histopathological section image. The method comprises the steps: carrying out the foreground segmentation of thehistopathological section image, and extracting a cell tissue area to obtain a foreground image; detecting different types of lesion regions in the foreground image by using a semantic segmentation network model based on a deep residual network structure and a multi-scale hole convolution structure; morphological post-processing is carried out on the detected lesion areas of different types to remove fine connection between the lesion areas of different types, and holes are filled, so that contours of the lesion areas of different types are obtained; and optimizing the contours of different types of lesion regions by using a deformation model established in combination with the global shape information so as to complete automatic detection of the lesion regions in the whole pathological tissue slice image. According to the mode of combining the deep semantic segmentation network and the deformation model, priori knowledge and global information can be integrated, and the semantic segmentation accuracy is improved.

Description

technical field [0001] The invention belongs to the field of medical pathological image processing, and more specifically relates to a method and system for automatically detecting lesion regions in pathological tissue slice images based on a deep semantic segmentation network and a deformation model. Background technique [0002] Breast cancer is one of the most common malignant tumors in women, and the incidence of malignant tumors in women is extremely high. Clinically, early screening for breast cancer usually involves palpation and regular examinations, with initial detection using mammography or ultrasound imaging, and breast biopsy if testing indicates the possibility of malignant tissue growth. Pathologists distinguish normal tissue, benign and malignant lesions based on breast pathological images and perform prognosis assessment. Accurate detection and classification of pathological images is an important basis for pathologists to make the best treatment plan. But...

AI Technical Summary

Problems solved by technology

[0004] In view of the above defects or improvement needs of the prior art, the present invention provides a method and system for automatic detection of lesion areas in pathological tissue slice images, thereby solving the problem of a single lesion area in the existing detection of pathological images based on deep convolutional neural networks Identify technical issues with inconsistent results

Images