A case study on the use of Amazon visual ID facial recognition metadata in investigation

Deep neural networks have been part of many breakthroughs in computer graphics and vision research. In the context of visual content synthesis, deep learning models have achieved impressive performance in the image domain. However, adapting the successes of image synthesis models to the video domain has been difficult, arguably due to the lack of sufficiently strong inductive biases that encourage the models to capture the temporal-dynamic nature of video data. Inductive bias refers to the prior knowledge incorporated into the learning models to explicitly drive the learning process toward the solutions that capture meaningful structures from data, which is critical to help the model generalize beyond the training data. Successful deep neural network architectures, such as convolutional neural networks (CNN), while effective in representing image data thanks to the spatial inductive bias, often lack the inductive biases relating to the dynamic nature of videos. Mai argues that designing such inductive biases can benefit from the domain knowledge of video processing literature. Their primary motivation in this thesis is to demonstrate that the knowledge acquired from traditional computer vision and graphics literature can serve as effective inductive biases for designing deep learning models for video synthesis. This dissertation provides the initial steps toward verifying that insight via two case studies. In the first case study, Mai explored adapting the standard CNN architecture to perform video frame interpolation. Early CNN-based methods for frame generation followed the direct prediction approach, thus ineffective in learning to capture motion information. Inspired by traditional video frame interpolation techniques that established frame interpolation as a joint process of motion estimation and pixel re-sampling, Mai presented the CNN-based frame interpolation framework that incorporated such insight into the synthesis model via the novel AdaConv layer. That serves as a functional inductive bias and enables the first deep learning model for high-quality video frame interpolation. In the second case study, Mai explored adapting the recent Implicit Neural Representation (INR) to a novel motion-adjustable video representation. Viewing modern INR frameworks as a form of non-linear transform from a frequency domain to the image domain, and inspired by the success of phase-based motion modelling in the classical computer vision literature, they presented a simple modification to the standard image-based INR model that allows for not only video reconstruction but also a variety of motion editing tasks.

Computer vision (CV) is a branch of artificial intelligence (AI) that enables machines to understand visual input. The recent rise of deep learning (DL) has empowered CV significantly, leading to well-established applications such as autonomous vehicles, medical diagnosis and facial recognition. These new capabilities extend to the manufacturing sector, however they have not been widely adopted to monitor processes in the steel industry due to challenges related to harsh environmental conditions such as poor lighting, heat distortion, dust particles and vibrations. As a result, existing datasets are limited and advances have predominantly been evaluated within research settings but not real-world settings. Therefore, this project investigates the application of CV for monitoring steel production processes and how integration impacts state-of-the-art technology. This work aims to produce CV systems capable of monitoring different processes and utilise them to draw valuable real-world insights for industry. Also, it aims to investigate how these systems, and CV as a whole, can enhance the efficiency, quality and sustainability of steel manufacturing. This research involves the development of CV models tailored to three processes: ladle pouring, galvanising and gas stirring. In each case study, DL and traditional methods are used to monitor real or simulated production environments and extract useful information. Primary outcomes of this research include a foundation for monitoring ladle pouring to reduce emissions, a deployed system for quantifying zinc splatter occurring during galvanisation in real-time, and a tool for comparing the wear rate and stirring efficiency of different gas stirring approaches. Outcomes of this work highlight the revolutionary benefits of applying CV in production environments for process monitoring and control. By developing CV models for monitoring processes, overcoming harsh conditions typical in production environments, and drawing valuable insights from CV application, this work establishes a strong foundation for real-world implementation of CV in manufacturing.

Biometric systems are becoming more and more efficient due to increasing performance of algorithms. These systems are also vulnerable to various attacks. Presentation of falsified identity to a biometric sensor is one the most urgent challenges for the recent biometric recognition systems. Exploration of specific properties of thermal infrared seems to be a comprehensive solution for detecting face presentation attacks. This letter presents outcome of our study on detecting 3D face masks using thermal infrared imaging and deep learning techniques. We demonstrate results of a two-step neural network-featured method for detecting presentation attacks. Full Text: PDF ReferencesS.R. Arashloo, J. Kittler, W. Christmas, "Face Spoofing Detection Based on Multiple Descriptor Fusion Using Multiscale Dynamic Binarized Statistical Image Features", IEEE Trans. Inf. Forensics Secur. 10, 11 (2015). CrossRef A. Anjos, M.M. Chakka, S. Marcel, "Motion-based counter-measures to photo attacks inface recognition", IET Biometrics 3, 3 (2014). CrossRef M. Killioǧlu, M. Taşkiran, N. Kahraman, "Anti-spoofing in face recognition with liveness detection using pupil tracking", Proc. SAMI IEEE, (2017). CrossRef A. Asaduzzaman, A. Mummidi, M.F. Mridha, F.N. Sibai, "Improving facial recognition accuracy by applying liveness monitoring technique", Proc. ICAEE IEEE, (2015). CrossRef M. Kowalski, "A Study on Presentation Attack Detection in Thermal Infrared", Sensors 20, 14 (2020). CrossRef C. Galdi, et al, "PROTECT: Pervasive and useR fOcused biomeTrics bordEr projeCT - a case study", IET Biometrics 9, 6 (2020). CrossRef D.A. Socolinsky, A. Selinger, J. Neuheisel, "Face recognition with visible and thermal infrared imagery", Comput. Vis Image Underst. 91, 1-2 (2003) CrossRef L. Sun, W. Huang, M. Wu, "TIR/VIS Correlation for Liveness Detection in Face Recognition", Proc. CAIP, (2011). CrossRef J. Seo, I. Chung, "Face Liveness Detection Using Thermal Face-CNN with External Knowledge", Symmetry 2019, 11, 3 (2019). CrossRef A. George, Z. Mostaani, D Geissenbuhler, et al., "Biometric Face Presentation Attack Detection With Multi-Channel Convolutional Neural Network", IEEE Trans. Inf. Forensics Secur. 15, (2020). CrossRef S. Ren, K. He, R. Girshick, J. Sun, "Proceedings of IEEE Conference on Computer Vision and Pattern Recognition", Proc. CVPR IEEE 39, (2016). CrossRef K. He, X. Zhang, S. Ren, J. Sun, "Deep Residual Learning for Image Recognition", Proc. CVPR, (2016). CrossRef K. Mierzejewski, M. Mazurek, "A New Framework for Assessing Similarity Measure Impact on Classification Confidence Based on Probabilistic Record Linkage Model", Procedia Manufacturing 44, 245-252 (2020). CrossRef

The advent of Industry 4.0 has led to the widespread adoption of Industrial Internet of Things (IIoT) and Computer Vision technologies in manufacturing. However, handling the massive data generated by IIoT poses challenges. To address this, we propose a software and hardware architecture that combines edge computing with IIoT to enable computer vision applications on manufacturing floors. This research contributes to integrating edge computing and IIoT with Computer Vision technologies in manufacturing facilities, offering a cost-effective and scalable solution for machine vision applications. This approach benefits Small and Medium Manufacturers (SMMs) by utilizing low-cost hardware and open-source technologies, enabling them to harness Industry 4.0 advantages. Our approach involves IoT devices using Raspberry Pi microcomputers and USB webcams as clients and a Python-based edge node server software as the central component. Our test case study results show that a centralized edge-computing solution can offer several useability benefits and a significant performance increase over using a non-edge-enabled IIoT approach. The proposed architecture is tested, validated, and verified through a case study focused on the digitization of the readings of analog gauges. Our experimentation results can offer valuable insights and guidance for the broader adoption of edge-enabled IIoT solutions in manufacturing contexts.

The world's oceans are undergoing rapid changes due to climate change and other anthropogenic impacts, affecting marine species' distribution, abundance, and behavior. Traditional ecological monitoring methods struggle to keep pace with these transformations, especially in underwater habitats. Recently, computer vision techniques have emerged to enhance the efficiency of video and image-based underwater monitoring. These methods facilitate the detection and classification of objects in visual data, potentially streamlining the process of counting and classifying marine organisms. However, the adoption of computer vision in marine ecology has been slow, partly due to its inaccessibility to ecologists and the lack of easily adaptable tools for ecological monitoring.This study investigates the application and validation of computer vision techniques for monitoring underwater pelagic macrofaunal diversity, using a case study from coastal British Columbia. Over 9000 hours of underwater video were collected from four sites over 18 months, using mounted cameras programmed to record five minutes of video every hour. Due to the infrequency of organisms present in the videos and challenges with underwater visibility, we created a stepwise iterative screening process to sequentially refine video data and aid the image annotation process. This involved using unsupervised classifiers (e.g. ResNet-18) to assist in reducing the number of background (i.e. 'empty water') images shown to annotators. To address the scarcity of annotations for certain taxa, an ‘adjacency filter’ was employed to increase the number of annotated frames for rare species. A rigorous QAQC process ensured standardization and minimized inter-annotator bias. Finally, a supervised computer vision model (YOLOv8) was trained with approximately 240,000 annotations to assess the presence and abundance of marine species over 18 months in the area. This approach provided high-resolution temporal data on the diversity and abundance of pelagic fish and gelatinous zooplankton at these sites.Here, we detail the process of employing these computer vision techniques for long-term underwater ecological monitoring, emphasizing accessibility for ecologists. A stepwise method for adapting computer vision techniques to achieve biodiversity monitoring objectives is presented, highlighting the strengths and limitations of our approach. We address our work in the context of the key barriers facing computer vision in underwater ecology, and provide tools for researchers seeking to incorporate AI in image or video-based marine research. Finally, we propose future directions for integrating these technologies into new and existing monitoring programs, and suggest priority areas for future research to advance the use of computer vision in underwater ecological monitoring.

PurposeThis study provides a safety prewarning mechanism, which includes a comprehensive risk assessment model and a safety prewarning system. The comprehensive risk assessment model is capable of assessing nine safety indicators, which can be categorised into workers’ behaviour, environment and machine-related safety indicators, and the model is embedded in the safety prewarning system. The safety prewarning system can automatically extract safety information from surveillance cameras based on computer vision, assess risks based on the embedded comprehensive risk assessment model, categorise risks into five levels and provide timely suggestions.Design/methodology/approachFirstly, the comprehensive risk assessment model is constructed by adopting grey multihierarchical analysis method. The method combines the Analytic Hierarchy Process (AHP) and the grey clustering evaluation in the grey theory. Expert knowledge, obtained through the questionnaire approach, contributes to set weights of risk indicators and evaluate risks. Secondly, a safety prewarning system is developed, including data acquisition layer, data processing layer and prewarning layer. Computer vision is applied in the system to automatically extract real-time safety information from the surveillance cameras. The safety information is then processed through the comprehensive risk assessment model and categorized into five risk levels. A case study is presented to verify the proposed mechanism.FindingsThrough a case study, the result shows that the proposed mechanism is capable of analyzing integrated human-machine-environment risk, timely categorising risks into five risk levels and providing potential suggestions.Originality/valueThe comprehensive risk assessment model is capable of assessing nine risk indicators, identifying three types of entities, workers, environment and machine on the construction site, presenting the integrated risk based on nine indicators. The proposed mechanism, which adopts expert knowledge through Building Information Modeling (BIM) safety simulation and extracts safety information based on computer vision, can perform a dynamic real-time risk analysis, categorize risks into five risk levels and provide potential suggestions to corresponding risk owners. The proposed mechanism can allow the project manager to take timely actions.

Computer vision (CV) is rapidly advancing as a tool to make conservation science more efficient, for example, by accelerating the annotation of images from camera traps and aerial surveys. However, before CV can become a widely used approach, several core technology challenges need to be addressed by the CV community. Taking into consideration several case studies in CV where tremendous progress has been made since the emergence of deep learning, this chapter will introduce core concepts in CV, survey several areas where CV is already contributing to conservation, and outline key challenges for the CV community that will facilitate the adoption of CV in mainstream conservation practice.

This paper presents a face recognition system design based on computer vision and the specific realization method, including the hardware structure, face recognition specific algorithm, drive, and software design and implementation of the application. The system uses ARM9 chip as the core, the built-in Linux operating system, a dual camera with external USB interface, combined with the application, the embedded 3D face recognition method combining with software and hardware technology is proposed. Introduction The identity authentication technology based on biological characteristics have rapid development in recent years, which, identity verification based on facial features is the most natural and direct mean, so the computer face recognition technology is one of the most active and challenging fields. Especially the superiority of the face recognition technology in non-con environment and situation without shocking the human being detected has already more than the identification method, like fingerprint and iris, and break the bottleneck of 2D face recognition, eliminates the influence of illumination, facial expression, and a series of factors. The technology of 3D face recognition has become a hot topic being studied by many scholars nowadays. With the continuous development of computer vision technology, face recognition technology based on computer vision is also in constant improvement. Overall design According to the characteristics of computer vision and face recognition technology and its application idea in the field of testing, this paper defines the general idea of face recognition technology based on computer vision. The whole process of embedded 3D computer vision face recognition is accomplished by a computer, the main work process is: face detection, facial feature location and face modeling and 3D face recognition. The design of system hardware The device selection, circuit design, sealing mode and system efficiency, accuracy, physical security and anti-attack ability has a direct relationship in face recognition technology based on computer vision, its structure as shown in figure 1.

SummaryThe digitisation of plant collections is bringing large quantities of information into accessible electronic databases. However, in recent decades, traditional taxonomic work in collections has declined, meaning that more specimens are only determined to family or genus, particularly when lacking key identification structures. If unaddressed, large‐scale digitisation risks widening the gap between well‐studied species and those lacking data.Hyperspectral reflectance and computer vision are two emerging approaches for identifying species, but these have yet to be cross‐compared for herbarium‐based taxonomy. Using Lithocarpus species as a case study, we compared classification accuracy obtained from leaf reflectance spectra with computer vision (implemented via Pl@ntNet), a RGB (red, green, blue) image‐based approach known to work well on specimens presenting reproductive structures. In the spectral approach, we assessed how much data are needed to optimise classification accuracy, how many species could be discriminated between, and whether close relatives were more frequently confounded.We found that Lithocarpus herbarium specimens were accurately identified to species from relatively small spectral datasets. Despite not incorporating reproductive structures, this was only 14% less accurate than Pl@ntNet.We suggest these rapid, nondestructive leaf reflectance measurements, paired with computer vision, could fill identification gaps in collections, particularly for specimens lacking reproductive features.

Field Programmable Gate Arrays (FPGAs) are a new addition to the world of data center acceleration. While the underlying technology has been around for decades, their application in data centers slowly starts gaining traction. However, there are myriad problems that hinder the widespread application of FPGAs in the data center. The closed source tool chains result in vendor lock-in and unstable tool flows. The languages used to program FPGAs require different design processes which are not easily learned by software developers. Compared to commodity solutions using CPUs and GPUs, FPGAs are more expensive and more time consuming to develop for. All of this and more make FPGAs a tough sell to people in need of task acceleration. Nonetheless, FPGAs also offer an opportunity to develop faster accelerators with a smaller energy envelop for rapidly changing applications. This work presents a solution to FPGA abstraction using the TaPaSCo framework. TaPaSCo simplifies moving between different FPGA architectures and automates scaling of accelerators across a multitude of devices. In addition, the framework provides a homogenized way of interacting with the accelerators. This thesis presents applications where FPGAs offer many benefits in the data center. Applications such as Semi-Global Block Matching which are difficult to compute on CPUs and GPUs due to the specific data transfer patterns, can be implemented highly efficiently an FPGAs. The presented work achieves over 35x of speedup on FPGAs compared to implementations of GPUs. FPGAs can also be used to improve network efficiency in the data center by replacing central network components with smart switches. The work presented here achieves up to 7x speedup over a classical distributed software implementation in a hash join scenario. Furthermore, FPGA can be used to bring new storage technologies into the data center by providing highly efficient consensus services right inside the network. The presented work shows that fetching pages remotely using a FPGA accelerated consensus system can be done as fast as 10us over the network which is only 55% of a conventional solution. These results make non-volatile network storage solutions as replacement for main memory viable. Lastly, this thesis presents a way of simulating parts of a brain with a very high level accuracy using FPGA. The spiking neural networks employed in the accelerator can benefit the research of brain functionality. The accelerator is capable of handling tens of thousands of neurons with a strict real time requirement of 50us per simulation step.

Human action is a visually complex phenomenon. Visual representation, analysis and recognition of human actions has become a key focus of research in computer vision, artificial intelligence, robotics and other related scientific disciplines. Various applications of automated action recognition include but not limited to intelligent health care monitoring, smart-homes, content based video search, animation and entertainment, human-computer interaction and intelligent video surveillance. The main focus of all these application areas surrounds a fundamental question: Given a human subject doing something in the field of sensory input, what is the person doing? If machine is able to correctly answer this question, it can greatly benefit computer vision system development and practical usage. However, machine recognition of human action is a daunting task due to complex motion dynamics, anthropometric variations, occlusion and high dependency over camera viewpoint. In this thesis, we exploit the importance of rich visual cues from human actions and utilize them to propose valuable solutions to human action recognition. The important problem of view-invariance under viewpoint variations is taken as a case study. We collect and explore these visual cues from geometrical relationships, spatio-temporal patterns and features, frequency domain signal analysis, contextual associations of actions and derive action representations for machine recognition. Actions are known as spatio-temporal patterns and temporal order plays an important role in their interpretations. We, therefore, explore invariance property of temporal order of actions during action execution and utilize it for devising a new view-invariant action recognition approach. We apply order constraint and feature fusion on local spatiotemporal features. These features are representation of choice for action recognition due to their computational simplicity, robustness to occlusion and minor view-point changes. We introduce STOPs (spatio-temporal ordered packets) that combine discriminative characteristics of multiple features for better recognition performance. In addition, we introduce spatio-temporal ordering constraint that removes discrepancy of orderless formation of bag-of-feature framework for action recognition. Furthermore, to deal with limitations of feature based approaches, we explore multiple view geometry which has alleviated various complex problems in computer vision. We thoroughly study applications of static and multi-body flow fundamental matrix in context of relating across-view information. We introduce spatio-temporally consistent dense optical flow to avoid explicit manual human body landmark point detection and explicit point correspondences. We employ rank constraint to derive novel tracking and training-free action similarity measures across viewpoint variations. Next, we investigate that despite the considerable success of geometrical techniques, computational complexity due to dense optical flow calculations plays a hindering role. Therefore, we study and track frequency domain analysis of action sequences. It leads toward the derivation of spatio-temporal correlation filters that use frequency domain filtering to give fast and efficient solutions to action recognition. However, these filters are originally view-dependent solutions. To achieve this objective, view clustering is explored that extends frequency domain techniques to achieve view-invariance. Contextual information is another important cue for interpreting human actions especially when actions exhibit interactive relationships with their context. These contextual clues become even more crucial when videos are captured in unfavorable conditions like extreme low light nighttime scenarios. We, therefore, take case study of night vision and present contextual action recognition at nighttime. We discover that context enhancement is imperative in such challenging multi-sensor environment to achieve reliable action recognition which leads us to develop novel context enhancement techniques for night vision using multi-sensor image fusion. Extensive experimentation on well-known action datasets is performed and results are compared with the existing action recognition approaches in literature. The research findings in this thesis greatly encourage the exploitation of spatia-temporal visual cues for deriving novel action recognition approaches and increasing their performance.

Augmented reality (AR) is a groundbreaking technology that fully immerse the user in a mixed 'reality' where the real and virtual coexist in a unique manner. Integrating more artificial intelligence (AI) and computer vision into AR devices can greatly improve user input and provide a whole novel interface. Through AI technology, such as gesture recognition, object tracking, and face recognition, AR systems can offer more intuitive and engaging interactions. An AR system featuring such improved AI technologies can process real-time data while having the contextual awareness to respond to users' input and the surrounding environment on the fly. It can also provide narrative integration, character development, and dynamic environments to users, thereby enabling them to have a more personalised and meaningful experience. This paper examines the evolution of new media by discussing the possibilities of using AI and computer vision in AR devices to create personalised experiences, all the while critically looking at technical challenges and the opportunities they present to the field of future research and development. It also looks at case studies across different sectors, such as education, training, tourism, gaming, retail, and aviation to justify the potential of future development of AI-enhanced AR.

Product quality is one of the primary criteria considered by customers when choosing an item. Additionally, it is an essential factor for companies to stand out in a highly competitive market. In the Manaus Industrial Hub (PIM), in a machine used for producing safety belts, defect detection is a crucial stage in the production process. To enhance this task, Artificial Intelligence (AI) was implemented, standing out for its high efficiency in analyzing and processing data in industrial environments. The data was captured in image format by a camera, and using Deep Learning (DL) techniques, an intelligent algorithm capable of detecting faults was developed. Due to its autonomous learning capability and ability to identify and characterize defects, this algorithm represents the future of automated inspection. It has already achieved significant success in applications such as object identification and classification, facial recognition, and fault diagnostics. Given this context, the aim of this study is to propose an ideal solution to minimize failures in the production process of safety belts. The proposal seeks to automate the currently manual step using the concept of computer vision with AI, ensuring greater efficiency and reliability in the production process. Materials and Methods: The research, development, and application of AI with the algorithm in the case study were conducted in the R&D laboratory of the company located in the Manaus Industrial Hub (PIM). The project utilized product inputs, a camera equipped with a lens for capturing images, and a computer for data storage and algorithm development. Results: The application of AI in this environment uses computer vision systems to process image data. For this, a program was developed in Python with the PySimpleGUI library. The trained model was evaluated based on loss and accuracy metrics on the test set, achieving values of 0 and 100%, respectively. During testing, new belts were used, reaching 100% accuracy in the results. Conclusion: The proposed model showed excellent results. With data processed by AI using Deep Learning (DL) techniques, real-time inspection of the belts was achieved. Additionally, the network achieved perfect accuracy and recall in all tests conducted on the belts, demonstrating the effectiveness of the solution

Abstract-Global partnerships are essential for addressing pressing challenges such as climate change, humanitarian crises, and sustainable development, yet their effectiveness is often compromised by limited transparency, accountability gaps, and weak monitoring mechanisms. Traditional oversight methods, including audits and self-reported data, struggle to meet the demands of increasingly complex cross-border collaborations. In this context, computer vision emerges as a transformative technology capable of delivering automated, objective, and scalable analysis of visual data to ensure greater trust and accountability. This chapter explores the role of computer vision in enhancing transparency within global partnerships, beginning with a review of existing digital tools and continuing with a technical overview of vision-based systems and their applications in compliance monitoring, ethical sourcing, and equitable resource distribution. A proposed framework for integrating computer vision into partnership ecosystems is presented, supported by case studies from humanitarian aid, environmental governance, and global supply chains. The discussion also addresses challenges such as algorithmic bias, surveillance risks, and data privacy, while emphasizing future opportunities through integration with blockchain, IoT, and explainable AI. Ultimately, the chapter argues that computer vision, when embedded within robust ethical and policy frameworks, can act as a critical enabler of transparency, accountability, and trust in global cooperation. Keywords: Computer Vision, Global Partnerships, Transparency and Accountability, Artificial Intelligence in Governance, Supply Chain Monitoring, Ethical AI, Blockchain Integration, Digital Trust Frameworks, Automated Compliance Monitoring, Sustainable Development Goals.

A knowledge-based component library for high-level computer vision tasks