Projection full-convolution network three-dimensional model segmentation method based on fusion of multi-view-angle features

Abstract

The invention discloses a projection full-convolution network three-dimensional model segmentation method based on fusion of multi-view-angle features, and the method comprises the steps: 1, collecting data of a data set inputted into a three-dimensional network model; 2, carrying out the sematic segmentation of a model projection rendering map through a FCN (full-convolution network) integrated with multi-view-angle features, and obtaining the probability that the pixels of the projection rendering map of a model in all view-point directions are predicted as labels; 3, carrying out the back projection of a projection rendering map sematic segmentation probability graph of the model in each view-point direction, employing the maximum view-angle pooling, and obtaining the probability that amodel patch is predicted as each label; 4, carrying out the optimization through employing a Graph Cut image segmentation algorithm, and obtaining a final prediction label of the model patch.

Description

Technical field [0001] The invention belongs to the field of computer image processing and computer graphics, and particularly relates to a projection full convolutional network three-dimensional model segmentation method based on fusion of multi-view features. Background technique [0002] In recent years, with the emergence of more and more 3D modeling software, and depth sensors, such as Kinect, etc., have been widely used on platforms that collect depth data. Three-dimensional model data has exploded on the Internet. 3D The model also appeared in a large number of manifestations, such as point clouds, voxels, and patches. This trend makes the analysis of 3D models a hot research field. At present, the research in the field of image analysis has achieved fruitful results, and the introduction of the deep learning framework has further improved the effect. However, convolution operations on 2D images cannot be directly applied to 3D models, making it difficult to apply deep l...

AI Technical Summary

Problems solved by technology

[0007] 4. Analysis methods must be robust to noise, downsampling, and the diversity of similar models

However, the choice of viewing angle of this method is too time-consuming, and the optimization of CRF training at the end prolongs the training time of the entire framework, which is less practical.

Images