Point cloud completion is a generation and estimation issue derived from the partial point clouds, which plays vital role in applications 3D computer vision. The progress of deep learning (DL) has impressively improved capability robustness completion. However, quality completed clouds still needed to be further enhanced meet practical utilization. Therefore, this work aims conduct comprehensive survey on various methods, including point-based, convolution-based, graph-based, generative...

Artificial intelligence (AI) and 5G system have been two hot technical areas that are changing the world. On deep convergence of computing communication, networking systems AI (NSAI) is presenting a paradigm shift, where distributed becomes immersive in all elements network, i.e., cloud, edge, terminal devices, which make virtually operating as system. other hand, by evolution communication systems, network becoming service-specific interweaved with AI, operates an system, enabling real-time...

Graph processing is an important analysis technique for a wide range of big data applications. The ability to explicitly represent relationships between entities gives graph analytics significant performance advantage over traditional relational databases. However, at the microarchitecture level, bounded by inefficiencies in memory subsystem single-machine in-memory analytics. This paper consists two contributions which we analyze and optimize hierarchy workloads. First, perform in-depth...

Existing anomaly detection (AD) approaches rely on various hand‐crafted representations to represent video data and can be costly. The choice or designing of representation difficult when faced with a new dataset without prior knowledge. Motivated by feature learning, e.g. deep leaning the ability directly learn useful model high‐level abstraction from raw data, authors investigate possibility using universal approach. objective is learning data‐driven for task AD relying representation. A...

As a prerequisite for autonomous driving, scene understanding has attracted extensive research. With the rise of convolutional neural network (CNN)-based deep learning technique, research on achieved significant progress. This paper aims to provide comprehensive survey learning-based approaches in driving. We categorize these works into four work streams, including object detection, full semantic segmentation, instance and lane line segmentation. discuss analyze according their...

As one of the important applications intelligent video surveillance, violent behaviour detection (VioBD) plays a crucial role in public security and safety. particular type recognition, VioBD aims to identify whether behaviours that occurred scene is aggressive, such as fighting assault. To comprehensively analyse current state predict future trend research, we survey existing approaches this work. First, briefly introduce basic principle challenges VioBD; Then, category according their...

Achieving rapid and accurate localization of winter jujubes in trees is an indispensable step for the development automated harvesting equipment. Unlike larger fruits such as apples, jujube smaller with a higher density serious occlusion, which obliges requirements identification positioning. To address issues, method using improved YOLOX-Nano network was proposed. First, dataset containing variety complex scenes, backlit, occluded, different fields view, established to train our model....

Next Generation Sequencing (NGS) technology has become an indispensable tool for studying genomics, resulting in exponentially growth of biological data. Booming data volume demands significant computational resources and creates challenges 'Sequence Alignment', which is the most fundamental application bioinformatics. Consequently, many researchers exploit both software hardware methods to accelerate widely used sequence alignment algorithm - Basic Local Alignment Search Tool (BLAST)....

Local binary pattern (LBP) is one of the most successful feature descriptors. However, LBP and its variants have not been as other descriptors in video anomaly detection (VAD). This because are mainly designed for spatial texture analysis. Although volume (VLBP) LBP-three orthogonal planes (LPB-TOP) capability describing dynamic texture, they seldom used VAD 1) both VLBP LBP-TOP more suitable natural scenes with rich textures, but sensitive to noise less 2) combination motion appearance only...

Distributed fiber optic vibration sensing system has been widely used in safety monitoring with distinct advantages, and the feature extraction classification methods of signals directly determine real-time performance reliability system. The existing research are unidimensional time consuming, which cannot balance goals high accuracy low consumption for systems. In this paper, we propose an efficient recognition framework fusing signal time-frequency features, called TFF-CNN, based on...

Corona Virus Disease 2019 (COVID-19) poses a significant threat to human health and safety. As the core of prevention control COVID-19, safety medical nursing personnel are extremely important, standardized use personal protective equipment can effectively prevent cross-infection. Due existence severe occlusion overlap, traditional image processing methods struggle meet demand for real-time detection. To address these problems, we propose ME-YOLO model, which is an improved model based on...

Linear algebra plays an important role in computer science, especially cryptography. Numerous cryptographic protocols and scientific computations are based on linear algebra, which can be reduced to some core problems, such as matrix multiplication, determinant the characteristic polynomial of a matrix. However, it is difficult carry out these expensive independently for resource-limited cloud users. Outsourced computation, service provided by computing, enables resources-constrained client...

Detecting abnormal events in crowded scenes is an important but challenging task computer vision. Contextual information useful for discovering salient scenes; however, it cannot be characterized well by commonly used pixel-based descriptors, such as the HOG descriptor. In this paper, we propose contextual gradients between two local regions and then construct a histogram of oriented gradient (HOCG) descriptor event detection based on gradients. The HOCG distribution sub-regions different...

The Rician noise formed in magnetic resonance (MR) imaging greatly reduced the accuracy and reliability of subsequent analysis, most existing denoising methods are suitable for Gaussian rather than noise. Aiming to solve this problem, we proposed fuzzy c-means adaptive non-local means (FANLM), which combined (NLM) with (FCM), as a novel method reduce study.The algorithm chose optimal size search window automatically based on variance was estimated by improved estimator median absolute...

We propose a novel local nearest neighbor distance (LNND) descriptor for anomaly detection in crowded scenes. Comparing with the commonly used low-level feature descriptors previous works, LNND has two major advantages. First, efficiently incorporates spatial and temporal contextual information around video event that is important detecting anomalous interaction among multiple events, while most existing only contain of single event. Second, compact representation its dimensionality...