Three-dimensional model classification method based on geometrical shape and LSTM neural network

Abstract

The invention relates to a three-dimensional model classification method based on a geometrical shape and an LSTM neural network. According to the inventionn, the method includes extracting geometrical shape feature vectors D1, D2, D3 and A3 of a three-dimensional model, taking the feature vectors as a vector sequence to serve as input to be transmitted to an LSTM neural network, and training a noise reduction auto-encocoder, wherein the encoding part of the noise reduction auto-encoder is an LSTM neural network, and the LSTM neural network performs feature extraction and aggregation on feature vectors to form global features for recognition and classification of a three-dimensional model; training the weight of the XGBoost classifier by using the global feature and the category label of the three-dimensional model in the training set; and utilizing the optimized LSTM neural network and the XGBoost classifier to classify the three-dimensional model in the test set. The invention has agood effect in the aspect of three-dimensional model classification.

Description

Technical field: [0001] The invention relates to a three-dimensional model classification method based on geometric shape and LSTM neural network, and the method has better application in three-dimensional model classification. Background technique: [0002] In recent years, with the rapid development of 3D modeling technology, imaging technology and computer vision, the types and quantities of 3D models have experienced explosive growth. How to effectively classify and manage these 3D models has become an urgent problem to be solved. Neural networks have been widely used in many fields such as speech recognition, machine translation, language modeling, character recognition and 3D model classification. There are some deficiencies in the classification of 3D models by traditional neural networks. When the 3D models are numerous and complex, the classification effect is not ideal. Geometry information can accurately reflect the differences between different 3D models and c...

AI Technical Summary

Problems solved by technology

[0003] In order to solve the problem of 3D model classification, the present invention discloses a 3D model classification method based on geometric shape and LSTM neural network

Images