Suspected micro-calcification point region automatic positioning method based on discriminant deep belief network

Abstract

The invention discloses a suspected micro-calcification point region automatic positioning method based on a discriminant deep belief network. The method comprises the following steps: 1) mammary gland X-ray image preprocessing: segmenting and enhancing a mammary gland region; 2) sample acquisition and preprocessing: segmenting the enhanced mammary gland image to obtain a sub-block image set for model training, and performing noise reduction and background removal processing on sub-blocks; 3) sub-block feature extraction and classification: constructing a discriminant deep belief network (DDBNs), training and finely adjusting a DDBNs model, and completing feature extraction and automatic classification of mammary gland sub-blocks; and 4) micro-calcification point region detection: inputting a mammary gland X-ray image to be detected, performing a series of preprocessing on the image, performing classification judgment on the sub-blocks by applying the trained optimal model, and markingsuspicious micro-calcification point regions according to a judgment result. According to the method, automatic detection and positioning of the suspicious lesion area can be completed, the false positive rate is effectively reduced while the high detection rate is obtained, and the detected calcification point area and the expert marking area have high consistency.

Description

technical field [0001] The invention belongs to the technical field of medical image processing, and in particular relates to a method for automatically locating areas of suspected microcalcifications based on a discriminative deep belief network. Background technique [0002] Breast cancer is one of the common malignant tumors that threaten women's physical and mental health. Microcalcification clusters in mammograms are an important sign of early breast cancer. Research data show that 30%-50% of diagnosed breast cancer patients have microcalcification clusters in early mammogram screening. The diagnosis of calcification based on mammography is the first choice for the diagnosis and screening of breast cancer today. However, due to the low contrast between the microcalcifications of some early breast cancers and the surrounding tissue, it is difficult for the human eye to identify them, which often leads to missed diagnosis. Diagnosis depends largely on radiologists, so it...

AI Technical Summary

Problems solved by technology

[0004] Because the microcalcifications in mammograms are very small, different in size, shape, and distrib

Images