Human Action Recognition (HAR) is a challenging domain in computer vision, involving recognizing complex patterns by analyzing the spatiotemporal dynamics of individuals’ movements in videos. These patterns arise in ...
详细信息
Matrix-vector multiplication (MVM) operations play an important role in applications such as data processing and artificial neural networks. To meet the growing demand for computing power, the photonic MVM processor p...
详细信息
Matrix-vector multiplication (MVM) operations play an important role in applications such as data processing and artificial neural networks. To meet the growing demand for computing power, the photonic MVM processor provides what we believe to be a new computing architecture. In this paper, we propose a reconfigurable parallel MVM (RP-MVM) processor. To further improve the parallel computing dimension, wavelength division multiplexing (WDM) and digital subcarrier multiplexing (DSM) technologies were first incorporated into the photonic MVM. Compared with the traditional WDM-MVM architecture, the parallelism of RP-MVM scheme is increased by N times, where N is the carrier number of DSM signal. Moreover, the input data channel can be dynamically adjusted without changing the hardware scale, which improves the flexibility of computing system. The simulation results show that the RP-MVM scheme can achieve parallel computing operations of eight MVMs, with a computing speed of 128 GOPs. For a random 6-bit resolution data sequence, the root mean square error (RMSE) of calculation results is on the order of 1E-3. In addition, for the image edge extraction task based on Roberts operator, this scheme can realize the parallel processing of four grayscale images. Therefore, the proposed scheme provides an alternative approach for realizing a highly parallel and reconfigurable large-scale photonic MVM architecture.
Quantum machine Learning (QML) promises the transformative potential in computer vision by utilizing quantum computing to facilitate faster high-dimensional data processing. In this paper, we will go through some of t...
详细信息
This paper presents a method for synthesizing 2D and 3D sensor data for various machinevision tasks. Depending on the task, different processing steps can be applied to a 3D model of an object. For object detection, ...
详细信息
In the last few years, the abundance of available plank-ton images has significantly increased due to advancements in acquisition system technology. Consequently, a growing interest in automatic plankton image classif...
详细信息
This paper investigates the optimization and deployment of YOLOv7 deep learning model on NVIDIA Jetson Nano, an AI-focused edge computing platform for object detection in various computer visionapplications. The work...
详细信息
Event cameras record sparse illumination changes with high temporal resolution and high dynamic range. Thanks to their sparse recording and low consumption, they are increasingly used in applications such as AR/VR and...
详细信息
Event cameras record sparse illumination changes with high temporal resolution and high dynamic range. Thanks to their sparse recording and low consumption, they are increasingly used in applications such as AR/VR and autonomous driving. Current top-performing methods often ignore specific event-data properties, leading to the development of generic but computationally expensive algorithms, while event-aware methods do not perform as well. We propose Event Transformer(+), that improves our seminal work EvT with a refined patch-based event representation and a more robust backbone to achieve more accurate results, while still benefiting from event-data sparsity to increase its efficiency. Additionally, we show how our system can work with different data modalities and propose specific output heads, for event-stream classification (i.e. action recognition) and per-pixel predictions (dense depth estimation). Evaluation results show better performance to the state-of-the-art while requiring minimal computation resources, both on GPU and CPU.
The significance of high-speed machinevision in scientific and technological fields is growing, especially with the era of Industry 4.0 technologies. There are several pattern-matching algorithms that have various in...
详细信息
The significance of high-speed machinevision in scientific and technological fields is growing, especially with the era of Industry 4.0 technologies. There are several pattern-matching algorithms that have various intriguing applications in ultralow-latency machinevisionprocessing. However, the low frame rate of image sensors-which usually operate at tens of hertz-fundamentally limits the processing rate. The paper will conceptualize and develop the computerized pattern recognition technique that can be applied to investigate light beam profiles and extract the desired information according to the purpose required in this case study. In the current work, the automatic detection and inspection of laser spots were designed to perform analysis and alignment for laser beam in comparison with the electron spot beam using the LabVIEW graphical programming environment, especially when the laser and electron beams overlap. This is one of the important steps for realizing the fundamental aim of test-FEL to produce short wavelengths with the second, third, and fifth harmonics at 131.5, 88, and 53 nm, respectively. The tentative version of the program achieved the elementary purpose, which fulfilled the accurate transversal alignment of the ultrashort laser pulses with the electron beam in the system of the FEL test facility at MAX-Lab, in addition to studying the beam's stability and jittering range. Copyright (C) 2024 The Authors.
Remote sensing scene categorization (RSSC) is a long-standing, vital, and complex issue in computer vision. It seeks to classify a scene into one of the predetermined scene groups by analysing the entire image. The ri...
详细信息
Remote sensing scene categorization (RSSC) is a long-standing, vital, and complex issue in computer vision. It seeks to classify a scene into one of the predetermined scene groups by analysing the entire image. The rise of large-scale datasets and the resurgence of deep learning-based methods, which directly learn potent feature representations from large amounts of raw data, have led to a lot of progress in representing and classifying RS scenes. Convolutional neural networks (CNN) are among the varieties of deep neural networks that have been the subject of the most research. Taking advantage of the swift increase in the amount of labelled samples and the major enhancements in the strength of processing units, CNNs research has advanced swiftly, producing state-of-the-art results on a number of applications. In this overview, we present a comprehensive evaluation of earlier published surveys and recent CNN-based approaches for RSSC. This study covers more than 100 significant works on scene categorization, including problems, benchmark datasets, and qualitative performance evaluation. In view of the results so far, this study concludes with a list of intriguing research opportunities.
image stabilization plays a crucial role in providing accurate and reliable visual information for machinevisionapplications. In maritime applications, such as unmanned ship navigation, where six degrees of freedom ...
详细信息
ISBN:
(纸本)9798350388350;9798350388343
image stabilization plays a crucial role in providing accurate and reliable visual information for machinevisionapplications. In maritime applications, such as unmanned ship navigation, where six degrees of freedom (DOF) motion and harsh maritime conditions prevail, the efficacy of image stabilization technology is vital for robust imageprocessing algorithms. This paper offers a comprehensive review of image stabilization techniques tailored for maritime environments, developed over the past two decades. We analyzed a total of 39 research articles on the subject, sourced from Web-of-Science, SCOPUS, and the Engineering Index databases, discussing potential research directions to address the limitations of current image stabilization methods, with special consideration for the unique requirements of ship-borne cameras. It provides an up-to-date overview of the techniques, limitations, and algorithms of ship-borne cameras for maritime applications, identifying current knowledge gaps and areas requiring further research. This review aims to guide the development of new technologies and methods to improve the performance of image stabilization systems in maritime contexts.
暂无评论