Method for detecting branch points of tree structure in digital image

Abstract

The invention discloses a branch point detection method for a tree structure in a digital image, and the method is a depth branch point detection model based on a two-stage cascaded convolutional network, i.e., a candidate region segmentation network and a false detection elimination network. The method comprises the following steps: firstly, extracting a sample with a fixed size from an originalimage to train a three-dimensional U-shaped segmentation network of an anisotropic convolution kernel, then inputting an image containing a tree structure into the trained segmentation network for segmentation to obtain branch point candidate regions, and taking each point of the candidate regions as a candidate point of a branch point; extracting three 3D image blocks of the candidate points according to three proportions, and calculating the maximum intensity projection of three views of each 3D image block to form nine corresponding 2D views; meanwhile, inputting the 2D views into the stacks of the five convolution layers respectively, finally, fusing the features, corresponding to the candidate points, of the 2D views after convolution, a final branch point detection result is obtained, and the method has the advantages of being low in calculation cost, low in false detection rate and high in detection efficiency.

Description

Technical field [0001] The present invention relates to the technical field of image processing, in particular to a method for detecting branch points of a tree structure in a digital image. Background technique [0002] In digital image research, the morphological reconstruction of tree-like structures (such as neurons, retinal blood vessels and bronchi) is very important; in neurobiology research, large-scale volume microscope image data sets for three-dimensional neuron reconstruction are important for understanding the brain The function of neural network is very important; in ophthalmology, the shape of the retinal vascular tree can provide important clinical information for the diagnosis of glaucoma, proliferative diabetic retinopathy and other diseases. In addition, the morphological reconstruction of the bronchus is of great significance in the research of various lung diseases, and can be used for quantitative research on bronchial diseases. [0003] So far, many research...

AI Technical Summary

Problems solved by technology

For example, DeepVesselNet uses point labels to train 3D CNN. Point labels specify the position of a single voxel, indicating the existence of vascular bifurcations. However, 3D CNN requires high memory and computing speed due to its huge amount of parameters. At the same time, due to overlapping slices Convolution calculations are repeated on the same voxel in , so voxel classification using 3D CNN will lead to inefficient detection tasks

Images