Hybrid feature CNN model for point cloud classification and segmentation

Abstract

This study proposes a hybrid feature convolutional neural network (HFCNN) model for the complete description of three-dimensional (3D) point cloud features. The HFCNN confers sensitivity to the local, global, and single-point properties simultaneously by a feature vector space expansion. Wherein, a pointwise convolutional network sub-model realises the extraction of the local features by using a pointwise convolutional operator to process point cloud data directly. To consider the global properties of the point cloud, a central-point radiation model is constructed as an input of the feature layer in a nonnetwork form. Meanwhile, the single-point behaviour is characterised by the solo point coordinate information in the network. Within the constructed solo-local-global feature space, i.e. the fusion of single point feature, local feature and global feature, the HFCNN model can handle 3D point cloud data with unstructured and unordered properties. The HFCNN can be directly applied to the point cloud classification and segmentation without the modification of the CNN structure and training procedure. The experimental results have shown the effectiveness of the proposed model in prediction of class labels and point-by-point labels.

Tongue Image Texture Classification Based on Image Inpainting and Convolutional Neural Network

“…After all the image processing, ResNet is used to train and test the classification model of tough and tender tongue inpainting image. Convolutional neural network (CNN) [ 19 ] is a class of deep neural networks containing convolutional computation, which is widely used in image classification. The ResNet draws on the advantages of traditional deep learning networks and introduces the residual learning method, which solves problems such as loss and loss of information in transmission; makes the whole network only need to learn the difference between input and output, simplifying the goal and difficulty of network learning; effectively solves the problem of gradient dissipation and gradient explosion that exists in deep networks, enabling the network to be as deep as possible [ 20 , 21 ].…”

“…The most common deep residual networks are mainly ResNet50 and ResNet101. The performance comparison of ResNet50 and ResNet101 on ImageNet validation dataset is shown in Table 1 [ 19 ].…”

CA‐PMG: Channel attention and progressive multi‐granularity training network for fine‐grained visual classification

“…The conventional image classification has many applications, such as ImageNet classification, gender classification [16], point cloud classification [17], railway track surface defect classification [18], hyperspectral image classification [19,20] and the other classification tasks [21][22][23]. These tasks belongs to coarse classification.…”

3D Model Classification Based on Bayesian Classifier with AdaBoost

“…Li et al [25] design a multiscale receptive field graph attention network with semantic features of local patch for point cloud, which captures abundant features of point cloud. Zhang et al [26] propose a hybrid feature CNN to describe features of 3D point cloud, which can handle 3D point cloud data with unstructured and unordered properties. Li et al [27] use X-Conv transformation to solve disorder of point clouds, and apply CNN to classify 3D point clouds.…”