An auxiliary detection method for automatically segmenting pulmonary nodules based on a deep convolutional neural network

Abstract

The invention discloses an auxiliary detection method for automatically segmenting pulmonary nodules based on a deep convolutional neural network. The method comprises the following steps: firstly, reading pulmonary CT image data; Preprocessing the lung CT image data, including image interpolation, denoising and normalization processing in sequence, so as to obtain a preprocessed image; Extractinga pulmonary parenchyma region and a blood vessel region from the obtained preprocessed image, and removing data outside the pulmonary parenchyma region and blood vessel region data in the preprocessed image to obtain a candidate region; Detecting and segmenting the candidate areas by using a deep convolutional neural network to obtain a plurality of independent candidate areas; Identifying the independent candidate region by using a deep convolutional neural network to obtain a nodule region; And finally carrying out fine tuning on the boundary of the nodule region to obtain the accurately segmented three-dimensional pulmonary nodule model. According to the method, pulmonary nodules can be effectively detected and segmented, and the boundaries of the nodules and other analogues can be well distinguished.

Description

technical field [0001] The invention belongs to the technical field of medical image detection methods, and in particular relates to an auxiliary detection method for automatically segmenting pulmonary nodules based on a deep convolutional neural network. Background technique [0002] Computer-aided diagnosis and treatment mainly refers to the use of imaging and medical image processing technology to use computers to analyze patient X-ray, CT, MRI, ultrasound and other physiological and biochemical data to assist doctors in finding lesions, diagnosing diseases, and planning treatment plans. Practice has proved that computer-aided diagnosis has made a great contribution to improving diagnostic accuracy, improving work efficiency, and reducing missed diagnoses. With the development of computer technology and artificial intelligence technology, computer-aided diagnosis is also becoming intelligent. [0003] Lung cancer is one of the malignant tumors with the fastest-growing mo...

AI Technical Summary

Problems solved by technology

In recent years, researchers have proposed some deep learning models for lung nodule detection and segmentation. Compared with previous methods, the effect has been significantly improved, but it also faces the following problems: the two-dimensional network cannot make good use of three-dimensional shape and texture information , it is difficult to correctly segment the three-dimensional boundary; the image and nodule features of the lung region have high complexity, and it is difficult to distinguish nodules from other similar objects (such as blood vessels)

Images