Search

With the gradual growth of deep learning in machine vision, efficient extraction of 3D point clouds becomes significant. The raw data of the 3D point cloud are sparse, disordered, and immersed in noise, which makes it difficult to classify and segment. Whether 3D point clouds can be classified and segmented or not, the local feature is an essential ingredient. Therefore, this paper proposes a GateNet-based PointNet++ network (G-PointNet++). G-PointNet++ extracts local features more accurately than PointNet++ by suppressing irrelevant features and emphasizing important features. Meanwhile, it refines the feature adaptively. Besides, the SENet and attention mechanism are introduced into PointNet++. G-PointNet++ was evaluated on the public ModelNet dataset, ShapeNet dataset, and S3DIS dataset, and its effectiveness in classification and segmentation tasks was verified. In the classification task, G-PointNet++ achieves an overall classification accuracy (OA) of 95.5% on ModelNet10 and 93.3% on ModelNet40. In the segmentation task, the mIoU of G-PointNet++ reaches 85.5% on ShapeNet. These experiments show that G-PointNet++ achieves better performance and saves more time than PointNet++, and its overall accuracy is higher than that of PCT network on ModeNet40. [ABSTRACT FROM AUTHOR]

The aim of this study is to improve the classification efficiency of advanced methods using a multilayered dictionary learning framework. This paper presents the new idea of "multilayered K-singular value decomposition (MLK-SVD)" dictionary learning as a multilayer method of classification. This method starts by building a sparse representation at the patch level and relies on a hierarchy of learned dictionaries to output a global sparse representation for the whole image. In this research using class labels of training data, the label information is associated with each dictionary item (columns of the dictionary matrix) to enforce discrimination in sparse codes during the dictionary learning process. Also, this algorithm instead of learning one shallow dictionary learned multiple levels of dictionaries. The proposed formulation of deep dictionary learning provides the basis to develop more efficient dictionary learning algorithms. It relies on a succession of sparse coding and pooling steps in order to find an efficient representation of the data for classification. The performance of the proposed method is evaluated on MNIST and CIFAR-10 datasets, and results show that this method can help in advancing the state of the art. [ABSTRACT FROM AUTHOR]