Development of an automated fruit classification system by using computer vision and deep learning.

Computer vision has advanced so far that machines now can think and see as we humans do. Especially deep learning has raised the bar of excellence in computer vision. However, the recent emergence of deep reinforcement learning is threatening to soar even greater heights as it combines deep neural networks with reinforcement learning along with numerous added advantages over both. This, being a relatively recent technique, has not yet seen many works, and so its true potential is yet to be unveiled. Thus, this chapter focuses on shedding light on the fundamentals of deep reinforcement learning, starting with the preliminaries followed by the theory and basic algorithms and some of its variations, namely, attention aware deep reinforcement learning, deep progressive reinforcement learning, and multi-agent deep reinforcement learning. This chapter also discusses some existing deep reinforcement learning works regarding computer vision such as image processing and understanding, video captioning and summarization, visual search and tracking, action detection, recognition and prediction, and robotics. This work further aims to elucidate the existing challenges and research prospects of deep reinforcement learning in computer vision. This chapter might be considered a starting point for aspiring researchers looking to apply deep reinforcement learning in computer vision to reach the pinnacle of performance in the field by tapping into the immense potential that deep reinforcement learning is showing.

Deep learning has emerged as a transformative technology in computer vision, enabling significant advancements in tasks such as image classification, object detection, segmentation, and anomaly detection across diverse domains, including healthcare, agriculture, robotics, and industrial automation. Despite these advancements, challenges related to model interpretability, data scarcity, generalization, computational demands, and real-time deployment remain significant barriers. This study aims to systematically review and analyze recent developments in deep learning techniques applied to computer vision, identify associated challenges and research gaps, and propose potential directions to enhance the efficiency, robustness, and applicability of these systems. A comprehensive literature search was conducted across multiple reputable databases, including ScienceDirect, SpringerLink, IEEE Xplore, MDPI, and Wiley Online Library, focusing on peer-reviewed articles published between 2018 and 2025. Thematic analysis and descriptive synthesis were applied to extract insights regarding deep learning architectures, application domains, datasets, key findings, and limitations. Results indicate that Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), Generative Adversarial Networks (GANs), Transformer-based architectures, and hybrid models have significantly advanced computer vision applications. However, issues such as interpretability, data scarcity, and computational complexity persist. Future directions include lightweight architectures, transfer learning, federated learning, explainable AI, and multi-modal approaches. In conclusion, while deep learning has substantially improved computer vision capabilities, addressing current limitations is essential for broader real-world adoption and multi-domain applicability.

Industrial storage tanks are widely used in petroleum, chemical industry, metallurgy and other process industries. The use of storage tanks are important in ensuring industrial safety production and product storage. Recently, concerns are rising over leaks and spills which leads to potential environmental and public health risks, for example, air pollution. Thus, there is a need to detect and monitor the tanks, which has become a regulatory requirement in most countries. The monitoring of industrial storage tanks in a city is of great practical significance to the planning and construction of the city. With the recent development of remote sensing technology, computer vision and image processing, it is possible to automatically detect industrial storage tanks from optical remote sensing images. This paper mainly studies the automatic detection of urban industrial storage tanks using deep learning based algorithm, aiming to help people manage and monitor the urban environment and resources. In this paper, we collected optical satellite remote sensing images of the city as a unit and created a city-level dataset including three cities of Guangdong province in south China by utilising satellite image data obtained from Google Earth imagery. To explore the effect of the deep learning object detection algorithm for the industrial storage tanks detection at the city-level, a deep learning model built with SSD (Single-Shot Detector) was trained based on our collected dataset. To further improve the detection accuracy of industrial storage tanks, Hough transform was used to reduce the false alarm in the deep learning results. The experimental results show that the combination of deep learning target detection model and Hough transform is effective in detecting of industrial storage tank on the collected city-level dataset and can achieve promising detection results.

[Objective] The field of computer vision has seen remarkable progress, largely due to the advancements in deep learning. These developments have revolutionized image recognition, interpretation, and application across numerous domains. This paper introduces a new framework designed to expand the potential of computer vision systems by harnessing the power of deep learning techniques. Deep neural networks are at the core of this new system, providing exceptional accuracy and reliability in tasks such as object recognition, image segmentation, and scene understanding. [Methodology] Furthermore, this framework offers a versatile platform for real-time image processing, paving the way for numerous applications in areas like industrial automation, medical diagnostics, and autonomous vehicles. This study comprehensively explores the architectural elements and methodologies that drive this innovative framework. It emphasizes the framework's technological capabilities, scalability, adaptability, and potential for broad adoption across industries seeking advanced computer vision solutions. [Results] The proposed model, Convolutional Neural Network-Feature Pyramid Network (CNN-FPN), demonstrates superior performance across all evaluated metrics for object detection compared to existing models. Specifically, it achieves the highest scores in Accuracy (57.2%), Recall (60.4%), Precision (94.1%), F1-Score (73.5%), and AUC (0.983). These results indicate that the proposed model offers superior performance and reliability in object detection tasks, demonstrating its potential for high-precision computer vision applications. [Conclusions] In conclusion, this innovative architecture represents a significant advancement in computer vision, enabled by the capabilities of deep learning. Our test findings demonstrate that compared to conventional algorithms, the enhanced CNN-FPN produced more accurate results.

Deep long-tailed learning, one of the most challenging problems in visual recognition, aims to train well-performing deep models from a large number of images that follow a long-tailed class distribution. In the last decade, deep learning has emerged as a powerful recognition model for learning high-quality image representations and has led to remarkable breakthroughs in generic visual recognition. However, long-tailed class imbalance, a common problem in practical visual recognition tasks, often limits the practicality of deep network based recognition models in real-world applications, since they can be easily biased towards dominant classes and perform poorly on tail classes. To address this problem, a large number of studies have been conducted in recent years, making promising progress in the field of deep long-tailed learning. Considering the rapid evolution of this field, this article aims to provide a comprehensive survey on recent advances in deep long-tailed learning. To be specific, we group existing deep long-tailed learning studies into three main categories (i.e., class re-balancing, information augmentation and module improvement), and review these methods following this taxonomy in detail. Afterward, we empirically analyze several state-of-the-art methods by evaluating to what extent they address the issue of class imbalance via a newly proposed evaluation metric, i.e., relative accuracy. We conclude the survey by highlighting important applications of deep long-tailed learning and identifying several promising directions for future research.

Deep learning (also known as deep structured learning, hierarchical learning, or deep machine learning) is a branch of machine learning based on a set of algorithms that attempt to model high-level abstractions in data by using a deep graph with multiple processing layers, composed of multiple linear and nonlinear transformations. Deep learning has been characterized as a class of machine learning algorithms with the following characteristics [257]: • They use a cascade of many layers of nonlinear processing units for feature extraction and transformation. Each successive layer uses the output from the previous layer as input. The algorithms may be supervised or unsupervised and applications include pattern analysis (unsupervised) and classification (supervised). • They are based on the (unsupervised) learning of multiple levels of features or representations of the data. Higher-level features are derived from lower-level features to form a hierarchical representation. • They are part of the broader machine learning field of learning representations of data. • They learn multiple levels of representations that correspond to different levels of abstraction; the levels form a hierarchy of concepts. These definitions have in common: multiple layers of nonlinear processing units and the supervised or unsupervised learning of feature representations in each layer, with the layers forming a hierarchy from low-level to high-level features. Various deep learning architectures such as deep neural networks, convolutional deep neural networks, deep belief networks (DBN), and recurrent neural networks have been applied to fields like computer vision, automatic speech recognition, natural language processing, audio recognition, and bioinformatics where they have been shown to produce state-of-the-art results on various tasks. In this chapter we start in Section 7.1 with an introduction, giving a brief history of this field, the relevant literature, and its applications. Then we study some basic concepts of deep learning such as convolutional neural networks, recurrent neural networks, backpropagation algorithm, restricted Boltzmann machines, and deep learning networks in Section 7.2. Then we illustrate three examples for Apache Spark implementation for mobile big data (MBD), user moving pattern extraction, and combination with nonparametric Bayesian learning, respectively, in Sections 7.3 through 7.5. Finally, we have summary in Section 7.6.

Deep learning has witnessed remarkable growth, demonstrating rapid advancements in multiple AI -related domains, such as computer vision and speech recognition. But due to constraints like computational resources, power consumption, and memory bandwidth, deploying deep learning models onto those edge devices or Internet of Things (IoT) applications is still challenging. Traditional cloud-based AI processing suffers from latency and security issues, rendering it unfit for real-time applications in autonomous systems, health monitoring, smart cities and factories. To overcome these shortcomings, Embedded Systems in Very Large-Scale Integration (VLSI) have proven to be a suitable solution to implementing deep learning models directly onto edge devices to decrease the dependence on cloud computing while also improving performance, efficiency, and security. Recent developments in hardware accelerators like as Field-Programmable Gate Arrays (FPGAs), Application Specific Integrated Circuits (ASICs), and Neuromorphic Processing Units (NPUs) made possible an efficient implementation of deep learning algorithms on the edges (embedded platforms). Approaches to model compression, such as quantization, pruning, knowledge distillation, and memory optimization, are fundamental to the development of compilation systems that lower computation costs while maintaining accuracy in deep neural networks. Additionally, energy-efficient architectures, such as systolic arrays and reconfigurable processing units, further enhance inference speed and power efficiency, making AI-enabled edge computing practical for real-time applications. Often, deep learning workloads can be served using VLSI-based accelerators, further improving the latency of IoT applications and saving bandwidth when integrating with complex deep learning frameworks. Edge devices with real-time AI capabilities are used in applications including smart surveillance, predictive maintenance, autonomous navigation, and biomedical signal processing. End devices that use embedded AI typically also have security advantages, where security risks that could be present elsewhere, such as in cloud-based AI processing, are reduced thanks to the embedded system's use of hardware-based keys and secure enclaves. Nonetheless, addressing issues like hardware-software co-design complexity, adaptability to morphing AI models, and generalization across varied edge computing scenarios are ongoing research efforts. This study investigates the utilization of VLSI-driven embedded systems for the effective execution of deep learning algorithms in edge equipment, as well as improvement approaches that can improve computational performance, power usage, and real-time processing abilities. This work contributes to the design of scalable, low-power, and high-performance AI systems for IoT by exploring a variety of hardware architectures and AI model compression techniques. The results will provide the basis for efficient and secure AI-enabled edge computing, contributing to the scalable adoption in smart environments, industrial automation, and new-generation IoT systems.

Deep learning has greatly increased the capabilities of "intelligent" technical systems over the last years [1]. This includes the industrial automation sector [1]-[4], where new data-driven approaches to, for example, predictive maintenance [2], computer vision [3], or anomaly detection [4], have resulted in systems more easily and robustly automated than ever before.

Deep reinforcement learning augments the reinforcement learning framework and utilizes the powerful representation of deep neural networks. Recent works have demonstrated the remarkable successes of deep reinforcement learning in various domains including finance, medicine, healthcare, video games, robotics, and computer vision. In this work, we provide a detailed review of recent and state-of-the-art research advances of deep reinforcement learning in computer vision. We start with comprehending the theories of deep learning, reinforcement learning, and deep reinforcement learning. We then propose a categorization of deep reinforcement learning methodologies and discuss their advantages and limitations. In particular, we divide deep reinforcement learning into seven main categories according to their applications in computer vision, i.e. (i) landmark localization (ii) object detection; (iii) object tracking; (iv) registration on both 2D image and 3D image volumetric data (v) image segmentation; (vi) videos analysis; and (vii) other applications. Each of these categories is further analyzed with reinforcement learning techniques, network design, and performance. Moreover, we provide a comprehensive analysis of the existing publicly available datasets and examine source code availability. Finally, we present some open issues and discuss future research directions on deep reinforcement learning in computer vision.

The chicken monitoring in closed cages is vital in welfare assessment and management of health factors. Computer vision can be relied upon for real-time automation of chicken health monitoring systems due to its non-invasive and invasive properties and its capacity to present a wide variety of information due to the development of information technologies. This article thoroughly overviews computer vision technology for poultry industry research. We recommend searching with the keywords 'computer vision' and 'chicken' or ‘broiler’ or 'health monitoring' or 'machine learning', or 'deep learning' were published between 2013 and early 2023 with open access provided by Diponegoro University only. All of the chosen articles were manually examined and categorized according to their applicability to computer vision in a poultry farm. This article summarizes the most recent developments in chicken health monitoring techniques utilizing computer vision systems, i.e., machine learning-based and deep learning-based systems. Prior to the successful implementation of this technology in the poultry industry, this article concludes by emphasizing the future work and significant challenges that must be addressed by researchers in the field of chicken health monitoring to guarantee the quality of this technology.

The competence of deep learning for the automation and manufacturing sector has received astonishing attention in recent times. The manufacturing industry has recently experienced a revolutionary advancement despite several issues. One of the limitations for technical progress is the bottleneck encountered due to the enormous increase in data volume for processing, comprising various formats, semantics, qualities and features. Deep learning enables detection of meaningful features that are difficult to perform using traditional methods. The book takes the reader on a technological voyage of the industry 4.0 space. Chapters highlight recent applications of deep learning and the associated challenges and opportunities it presents for automating industrial processes and smart applications. Chapters introduce the reader to a broad range of topics in deep learning and machine learning. Several deep learning techniques used by industrial professionals are covered, including deep feedforward networks, regularization, optimization algorithms, convolutional networks, sequence modeling, and practical project methodology. Readers will find information on the value of deep learning in applications such as natural language processing, speech recognition, computer vision, online recommendation systems, bioinformatics, and videogames. The book also discusses prospective research directions that focus on the theory and practical applications of deep learning in industrial automation. Therefore, the book aims to serve as a comprehensive reference guide for industrial consultants interested in industry 4.0, and as a handbook for beginners in data science and advanced computer science courses.

Recent developments in machine learning, particularly deep learning, have accelerated due to the combined innovations in computational throughput and algorithmic advancements. Simultaneously, many traditional systems (i.e., engineering systems, manufacturing systems) are also transforming into cyber-physical systems (CPS), with hardware and software components operating in tandem due to the declining cost of digitization. This transformation has created an abundance of opportunities to adapt deep learning approaches to optimize various aspects of these systems by leveraging the data these CPS generates. While deep learning has found successful adoption in areas such as computer vision and natural language processing, the potential benefits of applying deep learning to engineering design in CPS are still mostly untapped due to several existing challenges. Some of the challenges that impede the widespread adoption of deep learning for engineering design in CPS are (i) the existence of combinatorially-large design space, (ii) the ability to incorporate design constraints and domain-specific knowledge, (iii) the availability and cost of high-quality data acquisition, iv) identification of possible failure modes and v) robustifying the designs. While these challenges exist in conventional engineering design, methods to resolve or circumvent these challenges have been well developed based on years of experience. Addressing these challenges for deep learning used for engineering design in CPS will enable the development of deep learning-based data-driven methods to further optimize and support the engineering design processes. This dissertation includes multiple case studies of engineering design in CPS and develops various deep learning-based frameworks that address the challenges discussed above while simultaneously enhancing the design process. The first study involves the development of a supervised deep learning model that is applied to a niche additive manufacturing system for automated parameter selection and quality control. The next case study explores the use of deep reinforcement learning for a sequential design problem in a microfluidic device design example. Following that, a generative modeling framework for designing material microstructures will be presented. In the second part of the dissertation, threat models were developed for deep reinforcement learning-based controllers to identify potential failure modes in the controller designs for several robotic simulations. The final case study presents a possible approach for robustifying reinforcement learning-based controllers with application to power systems. Collectively, these case studies demonstrate the various ways in which deep learning can be adopted to circumvent existing challenges while improving different aspects of the engineering design problem.

Computational approaches to Explainable Artificial Intelligence: Advances in theory, applications and trends

Computer vision has been transformed by deep learning, which has made it possible for robots to comprehend and interpret visual data at previously unheard-of levels. This study explores deep learning for computer vision systems: future trends and upcoming technology. Looking ahead, several significant advancements are reshaping this profession. First off, a viable approach is to combine deep learning architectures with generative models like Variational Autoencoders (VAEs) and Generative Adversarial Networks (GANs). This synergy makes activities like picture generation, style transfer, and data augmentation easier and enables a more nuanced interpretation of visual data. Second, a key component of deep learning for computer vision is now recognized to be attention mechanisms. These techniques, which are based on human visual attention, allow models to concentrate on particular areas of a picture, improving efficiency and interpretability. Tasks such as object detection, image segmentation, and image captioning demonstrate the critical role of attention mechanisms. Moreover, computer vision is reaching new heights thanks to the combination of deep learning and other cutting-edge technologies. Real-time processing on devices is made possible by the combination of deep learning and edge computing, which lowers latency and improves privacy. Furthermore, new opportunities for applications in augmented reality, robotics, and immersive experiences are being created by the combination of deep learning and 3D vision technologies. In the deep learning space for computer vision, transfer learning and meta-learning are likewise becoming more and more popular [1]. These methods increase performance on a different but related job by utilizing pre-trained models or knowledge from one task. This method minimizes the need for large datasets for training in situations where labeled data is rare. Deep learning in computer vision raises ethical questions that are becoming more and more significant. Concerns about decision interpretability, algorithmic bias, and privacy are becoming more prevalent. Fairness, accountability, and transparency must be given top priority in this field's future developments in order to guarantee the appropriate use of these technologies. In conclusion, there are a lot of promising things that lie ahead for deep learning in computer vision. Advances in transfer learning, attention mechanisms, and the integration of generative models are all leading to more complex and effective vision systems. As these technologies advance, it is critical to address ethical issues in order to build trust and guarantee that deep learning has a good social impact. This study offers a thorough analysis of these recent developments, providing insight into the future direction of deep learning for computer vision.

Quality inspection of fruits and vegetables linked to food safety monitoring and quality control. Traditional chemical analysis and physical measurement techniques are reliable, they are also time-consuming, costly, and susceptible to environmental and sample changes. Hyperspectral imaging technology combined with deep learning methods can effectively overcome these problems. Compared with human evaluation, automated inspection improves inspection efficiency, reduces subjective error, and promotes the intelligent and precise fruit and vegetable quality inspection. This paper reviews reports on the application of hyperspectral imaging technology combined to deep learning methods in various aspects of fruits and vegetables quality assessment. In addition, the latest applications of these technologies in the fields of fruit and vegetable safety, internal quality, and external quality inspection are reviewed, and the challenges and future development directions of hyperspectral imaging technology combined with deep learning in this field are prospected. Hyperspectral imaging combined with deep learning has shown significant advantages in fruit and vegetable quality inspection, especially in improving inspection accuracy and efficiency. Future research should focus on reducing costs, optimizing equipment, personalizing feature extraction, and model generalizability. In addition, the development of lightweight models and the balance of accuracy, the enhancement of the database and the importance of quantitative research should also be brought to attention. These efforts will promote the wide application of hyperspectral imaging technology in fruit and vegetable inspection, improve its practicability in the actual production environment, and bring important progress for food safety and quality management.