Three-dimensional image classification method, apparatus and device, and computer readable storage medium

Abstract

The embodiment of the invention discloses a three-dimensional image classification method and device, equipment and a computer readable storage medium. The method comprises the following steps: acquiring a three-dimensional image and a plurality of text tags; projecting the three-dimensional image to a plurality of visual angles to obtain a plurality of visual angle depth images; coding the depth images of the plurality of visual angles and the plurality of text labels respectively to obtain image features corresponding to the plurality of visual angles and a plurality of text features; and calculating similarity according to the image features corresponding to the plurality of visual angles and the plurality of text features to obtain first text similarity. By converting an object classification task into an image-text matching task, the method is not limited by a fixed text label set, and has generalization ability for objects and concepts which are not seen. The first text similarity represents the similarity between the category corresponding to the object in the three-dimensional image and the plurality of text labels, and the category of the object in the three-dimensional image is determined according to the first text similarity, so that the classification accuracy of the three-dimensional image is improved.

Description

technical field [0001] The present application relates to the field of computer technology, and in particular to a three-dimensional image classification method, device, equipment and computer-readable storage medium. Background technique [0002] With the continuous development of network technology, more and more fields need to use classified 3D images, for example, machine learning model training, classification and recognition of a large number of 3D images, etc. Taking machine learning model training as an example, before recognizing the 3D images to be classified, it is necessary to provide a sample set of classified 3D images for training of the machine learning model. [0003] In the prior art, when classifying a 3D image, a pre-trained machine learning model is used to realize the detection and recognition process of the 3D image. When training a machine learning model, it is necessary to collect a large number of 3D image samples and text label sets corresponding ...

AI Technical Summary

Problems solved by technology

[0004] However, the pre-trained machine learning model can only classify the objects used in the training process, that is, limited by the fixed text label set, it cannot classify new types of objects, which reduces the accuracy of 3D image classification. sex

Images