Point2SpatialCapsule: Aggregating Features and Spatial Relationships of Local Regions on Point Clouds using Spatial-aware Capsules.

In point cloud processing, efficient feature aggregation of local regions is essential for hierarchical representations of 3D shapes. However, most prior works directly aggregate local features by symmetric functions, which will result in the desertion of vital shape information. In this paper, we propose a novel aggregation module based on 3D attention mechanism, named Local Region Attention Aggregation, which can capture the local shape implied in irregular points by emphasizing informative features and suppressing unnecessary ones. Specifically, for each local region, our module first encodes point relations, and then sequentially infers two attention maps generated by 3D geometric relations and fusion of multi-scale features respectively. After that, we multiply encoded point features with the attention maps to refine features adaptively. Consequently, by aggregating the higher-quality local features, shape awareness can be enhanced. Extensive experiments on various challenging benchmarks verify our method achieves state-of-the-art performance.

Object based video retrieval with local region tracking

This study analyzes the effectiveness of the global (the whole face) and local (regions of eyes, nose, and mouth) features for face recognition. Features describing human faces are encoded in local ternary patterns. The two-class support vector machine is used as the supervised learning algorithm for training recognition models. In the recognition process, recognition modes based on the global features and local features are cascaded. For identifying a face image, the local features are used iteratively for filtering out candidates that can not be clearly identified by the global features, until the one with highest possibility is concluded. The experimental results show that cascading the recognition models of global and local features obtains better classification accuracy than the single classification process.

Point cloud completion using multiscale feature fusion and cross-regional attention

The image-recipe cross-modal retrieval task, which retrieves the relevant recipes according to food images and vice versa, is now attracting widespread attention. There are two main challenges for image-recipe cross-modal retrieval task. Firstly, a recipe’s different components (words in a sentence, sentences in an entity, and entities in a recipe) have different weight values. If a recipe’s different components own the same weight, the recipe embeddings cannot pay more attention to the important components. As a result, the important components make less contribution to the retrieval task. Secondly, the food images have obvious properties of locality and only the local food regions matter. There are still difficulties in enhancing the discriminative local region features in the food images. To address these two problems, we propose a novel framework named Dual Cross Attention Encoders for Cross-modal Food Retrieval (DCA-Food). The proposed framework consists of a hierarchical cross attention recipe encoder (HCARE) and a cross attention image encoder (CAIE). HCARE consists of three types of cross attention modules to capture the important words in a sentence, the important sentences in an entity and the important entities in a recipe, respectively. CAIE extracts global and local region features. Then, it calculates cross attention between them to enhance the discriminative local features in the food images. We conduct the ablation studies to validate our design choices. Our proposed approach outperforms the existing approaches by a large margin on the Recipe1M dataset. Specifically, we improve the R@1 performance by +2.7 and +1.9 on the 1k and 10k testing sets, respectively.

The direct extension of 2D image learning to three-dimensional space is 3D point cloud learning. Recently, point cloud learning has shown significant results in 3D shape estimation and semantic segmentation. Despite these advancements, fundamental problems in point cloud learning still pose significant challenges. These problems include representation learning, shape generation, shape segmentation, and shape matching. In this paper, we propose a cognitive self-attention based learning approach to aggregate global representation of 3D shapes from point cloud data. We also integrate 3D point data with a binary tree structure to build a point cloud generator. We further design a novel Generative Adversarial Network (GAN) architecture to generate point clouds resembling the ground truth that could be used for unsupervised learning of 3D shapes. Relying on a self-attention mechanism, our framework, called SAPCGAN, aggregates the global graph features to correct the structural information of 3D shapes in the latent space. Finally, we compare the performance of two gradient penalty methods used in stabilizing the training of our GAN system. We show that our framework has high training efficiency and can provide state-of-the-art results in 3D point cloud generation. The performance of our is demonstrated with both quantitative and qualitative experimental evaluations. Furthermore, the generated 3D point clouds can be segmented into their natural semantic parts, such as, for example the four legs of a chair, the wings of an air plane, or the four wheels of a car.

Dual attention-based deep learning network for multi-class object semantic segmentation of tunnel point clouds

We investigated the application of multiresolution global and local texture features to reduce false-positive detection in a computerized mass detection program. One hundred and sixty-eight digitized mammograms were randomly and equally divided into training and test groups. From these mammograms, two datasets were formed. The first dataset (manual) contained four regions of interest (ROIs) selected manually from each of the mammograms. One of the four ROIs contained a biopsy-proven mass and the other three contained normal parenchyma, including dense, mixed dense/fatty, and fatty tissues. The second dataset (hybrid) contained the manually extracted mass ROIs, along with normal tissue ROIs extracted by an automated Density-Weighted Contrast Enhancement (DWCE) algorithm as false-positive detections. A wavelet transform was used to decompose an ROI into several scales. Global texture features were derived from the low-pass coefficients in the wavelet transformed images. Local texture features were calculated from the suspicious object and the peripheral subregions. Linear discriminant models using effective features selected from the global, local, or combined feature spaces were established to maximize the separation between masses and normal tissue. Receiver Operating Characteristic (ROC) analysis was conducted to evaluate the classifier performance. The classification accuracy using global features were comparable to that using local features. With both global and local features, the average area, Az, under the test ROC curve, reached 0.92 for the manual dataset and 0.96 for the hybrid dataset, demonstrating statistically significant improvement over those obtained with global or local features alone. The results indicated the effectiveness of the combined global and local features in the classification of masses and normal tissue for false-positive reduction.

In this study, we investigated how shapes are classified based on local and global features by four representative convolutional neural networks (CNNs), i.e., AlexNet, VGG, ResNet and Inception. While the local features are based on simple components, such as orientation of line segment, the global features are based on the whole object, such as whether an object has a hole. For example, solid triangles and solid squares are differentiated by local features, solid circles and rings are differentiated by a global feature. Two sets of experiments were performed in this research. In the first experiment, we examined how the four CNNs pre-trained on ImageNet (with transfer learning) learned to differentiate the regular shapes (equilateral-triangles, squares, circles and rings). Our results showed that the pre-trained CNNs exhibited faster learning rates in the tasks discriminating the local features than in the tasks discriminating the global feature. However, the transfer learning of discriminating the global feature in regular shapes were better generalized to irregular shapes than the transfer learning of discriminating local features. In the second experiment, the CNNs were trained from scratch (with random weights initialization) to discriminate local and global features in regular and irregular shapes. Different from the transfer learning, the CNNs exhibited faster learning rates in discriminating the global feature than the local features. Similar to transfer learning, the CNNs exhibited excellent learning generalization to discriminating the global feature of irregular shapes, but poor learning generalization to discriminating the local features in the irregular shapes. The overarching goal of this research is to create a paradigm and benchmark to directly compare how the CNNs and primate visual systems process geometrical invariants. In contrast to the ImageNet approach which employs natural images to train CNNs, we employed the “ShapeNet” approach which features geometrical shapes with well-defined properties. The ShapeNet approach will not only help elucidate the strengths and limitations of CNN computation, but also provide insights into visual information processing in the primates.

Point clouds obtained from 3D scanners are often noisy and cannot be directly used for subsequent high-level tasks. In this article, we propose a novel point cloud optimization method capable of denoising and homogenizing point clouds. Our idea is based on the assumption that the noise is generally much smaller than the effective signal. We perform noise perturbation on the noisy point cloud to get a new noisy point cloud, called self-variation point cloud. The noisy point cloud and self-variation point cloud have different noise distribution, but the same point cloud distribution. We compute the potential commonality between two noisy point clouds to obtain a clean point cloud. To implement our idea, we propose a Self-Variation Capture Network (SVCNet). We perturb the point cloud features in the latent space to obtain self-variation feature vectors, and capture the commonality between two noisy feature vectors through the feature aggregation and averaging. In addition, an edge constraint module is introduced to suppress low-pass effects during denoising. Our denoising method does not take into account the noise characteristics, and can filter the drift noise located on the underlying surface, resulting in a uniform distribution of the generated point cloud. The experimental results show that our algorithm outperforms the current state-of-the-art algorithms, especially in generating more uniform point clouds. In addition, extended experiments demonstrate the potential of our algorithm for point clouds upsampling.

Deep learning models for point clouds have shown to be vulnerable to adversarial attacks, which have received increasing attention in various safety-critical applications such as autonomous driving, robotics, and surveillance. Existing 3D attack methods generally employ global distance losses to implicitly constrain the point-wise perturbations for optimization. However, these simple losses are quite difficult to accurately measure and restrict the proper 3D geometry as point clouds are highly structured. Although few recent works try to exploit additional shape-aware surface knowledge to globally constrain the point position, they still fail to preserve the detailed point-to-point geometric dependency in different local regions. To this end, in this paper, we propose a novel Multi-grained Geometry-aware Attack (MGA), which explicitly captures the local topology characteristics in different 3D regions for adversarial constraint. Specifically, we first develop multi-scale spectral local filter banks adapting to different 3D object shapes to explore potential geometric structures in local regions. Considering that objects may contain complex geometries, we then extend each filter bank into multi-layer ones to gradually capture the topology contexts of the same region in a coarse-to-fine manner. Hence, the focused local geometric structures will be highlighted in the coefficients calculated by the filtering process. At last, by restricting these coefficients between benign and adversarial samples, our MGA is able to properly measure and preserve the detailed geometry contexts in the whole 3D object with trivial perturbations. Extensive experiments demonstrate that our attack can achieve superior performance on various 3D classification models, with satisfying adversarial imperceptibility and strong resistance to different defense methods.

Point clouds generated from real-world scanning are always incomplete and ununiformly distributed, which would cause structural losses in 3D shape representations. Therefore, a learning-based method is introduced in this paper to repair partial point clouds and restore complete shapes of target objects. First, we sample several local regions of inputs, encode their features and fuse them with independently extracted global features. Second, we establish a graph to connect all local features together, and then implement convolution with multi-head attention on the graph. Graph attention mechanism enables each local feature vector to search across the regions and selectively absorb other local features based on their relationships in high-dimensional feature space. Third, we design a coarse decoder to collect cross-region features from the graph and generate skeletons of complete point clouds, and a folding-based decoder is leveraged to generate final point clouds with high resolution. Our network is trained on six categories of objects from the ModelNet dataset, its performance is compared with several existing methods, the results show that our network is able to generate dense complete point cloud with the highest accuracy.

Reconstructing the incomplete point cloud into the complete and uniform one is a fundamental task in 3D point cloud processing. Some existing studies have explored the use of deep learning networks and representation learning to implement point cloud completion, but the completed shapes still appear unrealistic and non-uniform. To address this issue, we propose the idea of simulating the process of point cloud completion with local-to-global reasoning (LGR). Motivated to achieve the fine LGR, a novel mutual information (MI) maximization-based similarity operation is proposed, which realizes the reconstruction from incomplete point cloud to complete one by maximizing MI between global features and the prior features from the same 3D object. We adopt the Jenson-Shannon MI estimator to maximize the MI between global features and the priors in specific dimensions, which effectively increases the similarity of global and prior features. Compared with other existing works and similarity operations, the superior results indicate the efficacy of the proposed method and its advantages over existing ones in both the synthetic and real-world datasets. Our source code is available at <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/wendydidi/MISO-PCN</uri> .

For the existing 3D point cloud model classification method PointNet only extracts global feature information, but for the problem of insufficient local feature information extraction, a network that can take into account the combination of global features and local features is proposed, so as to realize 3D point Classification of cloud models. This paper proposes a classification network based on local location information. First, the point cloud model is sampled in multiple groups to obtain the center point of each group of samples, and then the radius is set to obtain its spherical neighborhood, and the points in the neighborhood and their center points are used as local location features. In information fusion, the acquired local information features are used as high-dimensional features of the neighborhood space through feature mapping. By analogy, multiple sets of sampled local features are integrated to realize the classification of the point cloud model. The experimental results show that the classification network can achieve a classification accuracy of 92.5% on the ModelNet40 data set, which is higher than the current mainstream algorithm.

Three-dimensional (3D) point clouds have a wide range of applications in the field of 3D vision. The quality of the acquired point cloud data considerably impacts the subsequent work of point cloud processing. Due to the sparsity and irregularity of point cloud data, processing point cloud data has always been challenging. However, existing deep learning-based point cloud dense reconstruction methods suffer from excessive smoothing of reconstruction results and too many outliers. The reason for this is that it is not possible to extract features for local and global features at different scales and provide different levels of attention to different regions in order to obtain long-distance dependence for dense reconstruction. In this paper, we use a parallel multi-scale feature extraction module based on graph convolution and an upsampling method with an added multi-head attention mechanism to process sparse and irregular point cloud data to obtain extended point clouds. Specifically, a point cloud training patch with 256 points is inputted. The PMS module uses three residual connections in the multi-scale feature extraction stage. Each PMS module consists of three parallel DenseGCN modules with different size convolution kernels and different averaging pooling sizes. The local and global feature information of the augmented receptive field is extracted efficiently. The scale information is obtained by averaging the different pooled augmented receptive fields. The scale information was obtained using the different average pooled augmented receptive fields. The upsampling stage uses an upsampling rate of r=4, The self-attentive features with a different focus on different point cloud data regions obtained by fusing different weights make the feature representation more diverse. This operation avoids the bias of one attention, and each focuses on extracting valuable fine-grained feature information. Finally, the coordinate reconstruction module obtains 1024 dense point cloud data. Experiments show that the proposed method demonstrates good evaluation metrics and performance and is able to obtain better visual quality. The problems of over-smoothing and excessive outliers are effectively mitigated, and the obtained sparse point cloud is more dense.