Journal articles on the topic 'Deep learning, computer vision, safety, road scene understanding'

In the past decade, autonomous vehicle systems (AVS) have advanced at an exponential rate, particularly due to improvements in artificial intelligence, which have had a significant impact on social as well as road safety and the future of transportation systems. However, the AVS is still far away from mass production because of the high cost of sensor fusion and a lack of combination of top-tier solutions to tackle uncertainty on roads. To reduce sensor dependency and to increase manufacturing along with enhancing research, deep learning-based approaches could be the best alternative for developing practical AVS. With this vision, in this systematic review paper, we broadly discussed the literature of deep learning for AVS from the past decade for real-life implementation in core fields. The systematic review on AVS implementing deep learning is categorized into several modules that cover activities including perception analysis (vehicle detection, traffic signs and light identification, pedestrian detection, lane and curve detection, road object localization, traffic scene analysis), decision making, end-to-end controlling and prediction, path and motion planning and augmented reality-based HUD, analyzing research works from 2011 to 2021 that focus on RGB camera vision. The literature is also analyzed for final representative outcomes as visualization in augmented reality-based head-up display (AR-HUD) with categories such as early warning, road markings for improved navigation and enhanced safety with overlapping on vehicles and pedestrians in extreme visual conditions to reduce collisions. The contribution of the literature review includes detailed analysis of current state-of-the-art deep learning methods that only rely on RGB camera vision rather than complex sensor fusion. It is expected to offer a pathway for the rapid development of cost-efficient and more secure practical autonomous vehicle systems.

Recent computer vision and deep learning breakthroughs have improved road safety by automatically classifying traffic signs. This research uses CNNs to classify traffic signs to improve road safety. Autonomous vehicles and intelligent driver assistance systems require accurate traffic sign detection and classification. Using deep learning, we created a CNN model that can recognize and classify road traffic signs. This research uses a massive dataset of labeled traffic sign photos for training and validation. These CNN algorithms evaluate images and produce real-time predictions to assist drivers and driverless cars in understanding traffic signs. Advanced driver assistance systems, navigation systems, and driverless vehicles can use this technology to give drivers more precise information, improving their decision-making and road safety. Researcher optimized CNN model design, training, and evaluation metrics during development. The model was rigorously tested and validated for robustness and classification accuracy. The research also solves real-world driving obstacles like illumination, weather, and traffic signal obstructions. This research shows deep learning-based traffic sign classification can dramatically improve road safety. This technology can prevent accidents and enhance traffic management by accurately recognizing and interpreting traffic signs. It is also a potential step toward a safer, more efficient transportation system with several automotive and intelligent transportation applications. Road safety is a global issue, and CNN-based traffic sign classification can reduce accidents and improve driving. On filter 3, Convolutional Neural Network training accuracy reached 98.9%, while validation accuracy reached 88.23%.

Understanding road scenes is crucial to the safe driving of autonomous vehicles, and object detection in road scenes is necessary to develop driving assistance systems. Current object detection algorithms are not very good at handling complex road scenes, and public datasets do not always adequately represent city traffic. Using Improved Multi-Scale Adaptive Feature Fusion Technology (IMSAFFT), this work suggests a real-time traffic information identification method to fix the issues of low detection accuracy of road scenes and high false detection rates in panoramic video images. In addition, a semantic recognition algorithm for a road scene based on image data is suggested. This study introduces computer vision-based approaches, including colour and texture recognition, object detection, and scene context understanding using Deep Neural Networks (DNN). An increasing number of deeper stacked layers allows the deep neural network to learn more complicated high-level semantic features, and the features' quality improves with time. A learning rate adaptive adjustment technique has been utilized to make training more efficient. After that, this improved detector is used to identify vehicles in original road environments. The suggested technique surpassed traditional detectors in the experiments with a high accuracy rate and processing speed. It worked well in real-world traffic situations for detecting overlapping, multiple, distant, and small objects. The simulation outcomes illustrate that the recommended IMSAFFT model increases the accuracy ratio of 98.4%, target image detection ratio of 97.4%, traffic prediction rate of 96.5%, processing speed rate of 10.4% and F1-score ratio of 95.4% compared to other existing models.

<span lang="EN-US">Intelligent transportation systems (ITS) are among the most focused research in this century. Actually, autonomous driving provides very advanced tasks in terms of road safety monitoring which include identifying dangers on the road and protecting pedestrians. In the last few years, deep learning (DL) approaches and especially convolutional neural networks (CNNs) have been extensively used to solve ITS problems such as traffic scene semantic segmentation and traffic signs classification. Semantic segmentation is an important task that has been addressed in computer vision (CV). Indeed, traffic scene semantic segmentation using CNNs requires high precision with few computational resources to perceive and segment the scene in real-time. However, we often find related work focusing only on one aspect, the precision, or the number of computational parameters. In this regard, we propose RBANet, a robust and lightweight CNN which uses a new proposed balanced attention module, and a new proposed residual module. Afterward, we have simulated our proposed RBANet using three loss functions to get the best combination using only 0.74M parameters. The RBANet has been evaluated on CamVid, the most used dataset in semantic segmentation, and it has performed well in terms of parameters’ requirements and precision compared to related work.</span>

The increasing availability of video data, through existing traffic cameras or dedicated field data collection, and the development of computer vision techniques pave the way for the collection of massive data sets about the microscopic behavior of road users. Analysis of such data sets helps in understanding normal road user behavior and can be used for realistic prediction of motion and computation of surrogate safety indicators. A multilevel motion pattern learning framework was developed to enable automated scene interpretation, anomalous behavior detection, and surrogate safety analysis. First, points of interest (POIs) were learned on the basis of the Gaussian mixture model and the expectation maximization algorithm and then used to form activity paths (APs). Second, motion patterns, represented by trajectory prototypes, were learned from road users' trajectories in each AP by using a two-stage trajectory clustering method based on spatial then temporal (speed) information. Finally, motion prediction relied on matching at each instant partial trajectories to the learned prototypes to evaluate potential for collision by using computing indicators. An intersection case study demonstrates the framework's ability in many ways: it helps reduce the computation cost up to 90%; it cleans the trajectory data set from tracking outliers; it uses actual trajectories as prototypes without any pre- and postprocessing; and it predicts future motion realistically to compute surrogate safety indicators.

Abstract: This paper presents a novel approach that addresses the challenging task of real-time drivable road region extraction in computer vision. Semantic segmentation, which involves accurately identifying and segmenting objects in real-time data, is a complex problem. However, deep learning has proven to be a powerful technique for achieving semantic segmentation by automatically identifying patterns without the need for explicit programming. To tackle this task, the paper proposes a fusion of the YOLO algorithm and UNET architecture, leveraging their respective strengths. The YOLO algorithm enables high-speed object detection, while the UNET architecture provides advantages in global location utilization, contextual understanding, and performance, even with limited training samples. Importantly, the proposed method is lightweight, making it suitable for deployment on embedded systems with limited computational power. To optimize memory usage and capture context at different scales, the system employs dilated convolutions for efficient feature extraction. The algorithm exhibits exceptional performance in accurately segmenting irregular objects and handles diverse input data types, including images and videos, in real-time. Overall, this paper contributes significantly to the advancement of computer vision technologies and offers a valuable solution for real-time drivable road region extraction. Its potential applications include addressing driving challenges and enhancing safety in autonomous vehicles and intelligent transportation systems.

Intelligent transportation systems (ITS) are among the most focused research in this century. Actually, autonomous driving provides very advanced tasks in terms of road safety monitoring which include identifying dangers on the road and protecting pedestrians. In the last few years, deep learning (DL) approaches and especially convolutional neural networks (CNNs) have been extensively used to solve ITS problems such as traffic scene semantic segmentation and traffic signs classification. Semantic segmentation is an important task that has been addressed in computer vision (CV). Indeed, traffic scene semantic segmentation using CNNs requires high precision with few computational resources to perceive and segment the scene in real-time. However, we often find related work focusing only on one aspect, the precision, or the number of computational parameters. In this regard, we propose RBANet, a robust and lightweight CNN which uses a new proposed balanced attention module, and a new proposed residual module. Afterward, we have simulated our proposed RBANet using three loss functions to get the best combination using only 0.74M parameters. The RBANet has been evaluated on CamVid, the most used dataset in semantic segmentation, and it has performed well in terms of parameters’ requirements and precision compared to related work.

For smart mobility, autonomous vehicles, and advanced driver-assistance systems (ADASs), perception of the environment is an important task in scene analysis and understanding. Better perception of the environment allows for enhanced decision making, which, in turn, enables very high-precision actions. To this end, we introduce in this work a new real-time deep learning approach for 3D multi-object detection for smart mobility not only on roads, but also on railways. To obtain the 3D bounding boxes of the objects, we modified a proven real-time 2D detector, YOLOv3, to predict 3D object localization, object dimensions, and object orientation. Our method has been evaluated on KITTI’s road dataset as well as on our own hybrid virtual road/rail dataset acquired from the video game Grand Theft Auto (GTA) V. The evaluation of our method on these two datasets shows good accuracy, but more importantly that it can be used in real-time conditions, in road and rail traffic environments. Through our experimental results, we also show the importance of the accuracy of prediction of the regions of interest (RoIs) used in the estimation of 3D bounding box parameters.

Autonomous driving is a safety-critical application that requires a high-level understanding of computer vision with real-time inference. In this study, we focus on the computational efficiency of an important factor by improving the running time and performing multiple tasks simultaneously for practical applications. We propose a fast and accurate multi-task learning-based architecture for joint segmentation of drivable area, lane line, and classification of the scene. An encoder–decoder architecture efficiently handles input frames through shared representation. A comprehensive understanding of the driving environment is improved by generalization and regularization from different tasks. The proposed method learns end-to-end through multi-task learning on a very challenging Berkeley Deep Drive dataset and shows its robustness for three tasks in autonomous driving. Experimental results show that the proposed method outperforms other multi-task learning approaches in both speed and accuracy. The computational efficiency of the method was over 93.81 fps at inference, enabling execution in real-time.

The article discusses the importance of selfdriving cars to improve road safety and reduce the number of accidents caused by human error. Self-driving cars not only reduce human error but also help reduce driver fatigue. We further explore the use of computer vision in autonomous cars, with previous research relying on deep learning algorithms with LiDAR sensors which can be expensive. The authors propose a more cost-effective approach using simple computer vision algorithms such as color space transformation, Canny edge detection, and Hough line transformation to detect lane lines and steer the car accordingly. This approach requires less operational hardware and can be implemented using affordable boards like Raspberry Pi and Nvidia Jetson Nano. The article also highlights the reconstruction of a remote-controlled car that had a 95% accuracy using a certain set of parameters was a tool for understanding autonomous cars better.

Aerial scenes captured by UAVs have immense potential in IoT applications related to urban surveillance, road and building segmentation, land cover classification, and so on, which are necessary for the evolution of smart cities. The advancements in deep learning have greatly enhanced visual understanding, but the domain of aerial vision remains largely unexplored. Aerial images pose many unique challenges for performing proper scene parsing such as high-resolution data, small-scaled objects, a large number of objects in the camera view, dense clustering of objects, background clutter, and so on, which greatly hinder the performance of the existing deep learning methods. In this work, we propose ISDNet (Instance Segmentation and Detection Network), a novel network to perform instance segmentation and object detection on visual data captured by UAVs. This work enables aerial image analytics for various needs in a smart city. In particular, we use dilated convolutions to generate improved spatial context, leading to better discrimination between foreground and background features. The proposed network efficiently reuses the segment-mask features by propagating them from early stages using residual connections. Furthermore, ISDNet makes use of effective anchors to accommodate varying object scales and sizes. The proposed method obtains state-of-the-art results in the aerial context.

Extracting information related to weather and visual conditions at a given time and space is indispensable for scene awareness, which strongly impacts our behaviours, from simply walking in a city to riding a bike, driving a car, or autonomous drive-assistance. Despite the significance of this subject, it has still not been fully addressed by the machine intelligence relying on deep learning and computer vision to detect the multi-labels of weather and visual conditions with a unified method that can be easily used in practice. What has been achieved to-date are rather sectorial models that address a limited number of labels that do not cover the wide spectrum of weather and visual conditions. Nonetheless, weather and visual conditions are often addressed individually. In this paper, we introduce a novel framework to automatically extract this information from street-level images relying on deep learning and computer vision using a unified method without any pre-defined constraints in the processed images. A pipeline of four deep convolutional neural network (CNN) models, so-called WeatherNet, is trained, relying on residual learning using ResNet50 architecture, to extract various weather and visual conditions such as dawn/dusk, day and night for time detection, glare for lighting conditions, and clear, rainy, snowy, and foggy for weather conditions. WeatherNet shows strong performance in extracting this information from user-defined images or video streams that can be used but are not limited to autonomous vehicles and drive-assistance systems, tracking behaviours, safety-related research, or even for better understanding cities through images for policy-makers.

Enhancing the environmental perception of autonomous vehicles (AVs) in intelligent transportation systems requires computer vision technology to be effective in detecting objects and obstacles, particularly in adverse weather conditions. Adverse weather circumstances present serious difficulties for object-detecting systems, which are essential to contemporary safety procedures, infrastructure for monitoring, and intelligent transportation. AVs primarily depend on image processing algorithms that utilize a wide range of onboard visual sensors for guidance and decisionmaking. Ensuring the consistent identification of critical elements such as vehicles, pedestrians, and road lanes, even in adverse weather, is a paramount objective. This paper not only provides a comprehensive review of the literature on object detection (OD) under adverse weather conditions but also delves into the ever-evolving realm of the architecture of AVs, challenges for automated vehicles in adverse weather, the basic structure of OD, and explores the landscape of traditional and deep learning (DL) approaches for OD within the realm of AVs. These approaches are essential for advancing the capabilities of AVs in recognizing and responding to objects in their surroundings. This paper further investigates previous research that has employed both traditional and DL methodologies for the detection of vehicles, pedestrians, and road lanes, effectively linking these approaches with the evolving field of AVs. Moreover, this paper offers an in-depth analysis of the datasets commonly employed in AV research, with a specific focus on the detection of key elements in various environmental conditions, and then summarizes the evaluation matrix. We expect that this review paper will help scholars to gain a better understanding of this area of research.

Abstract Road accidents constitute one of the leading causes of death and injuries in the global scene, and hence the requirement for developing efficient and real-time accident detection systems is great. This paper considers a vehicle accident detection system using computer vision techniques powered by OpenCV. Such a system utilises live video feeds that are acquired through surveillance cameras for monitoring vehicular movement with regard to accident detection due to sudden changes in velocity, direction, or collision events. It introduces features, including motion tracking, object detection with pre-trained deep learning models, and real-time algorithms for anomaly detection. Frame-by-frame analysis is also introduced where rapid deceleration and sudden stops can be identified in the system before a collision is detected. Also, optical flow and contour detection are utilized to identify accident-prone behaviour of traffic patterns. When an accident is established, the system will notify the emergency services; hence, response times will be reduced and lives will be saved. The results of experiments carried out in simulated scenarios prove the accuracy of the proposed system in accident detection with minimal false positives. With this research, a scalable solution comes into the forefront that can be introduced to infrastructures of smart cities to help improve the safety of traffic systems and implement proactive mechanisms for response to emergencies.Keywords Vehicle accident detection, Real-time alert system,Crash detection ,Accident severity analysis, Real-time location tracking opencv, Emergency notification.

Purpose This study aims to analyze driver risks in the driving environment. A complete analysis of context aware assistive driving techniques. Context awareness in assistive driving by probabilistic modeling techniques. Advanced techniques using Spatio-temporal techniques, computer vision and deep learning techniques. Design/methodology/approach Autonomous vehicles have been aimed to increase driver safety by introducing vehicle control from the driver to Advanced Driver Assistance Systems (ADAS). The core objective of these systems is to cut down on road accidents by helping the user in various ways. Early anticipation of a particular action would give a prior benefit to the driver to successfully handle the dangers on the road. In this paper, the advancements that have taken place in the use of multi-modal machine learning for assistive driving systems are surveyed. The aim is to help elucidate the recent progress and techniques in the field while also identifying the scope for further research and improvement. The authors take an overview of context-aware driver assistance systems that alert drivers in case of maneuvers by taking advantage of multi-modal human processing to better safety and drivability. Findings There has been a huge improvement and investment in ADAS being a key concept for road safety. In such applications, data is processed and information is extracted from multiple data sources, thus requiring training of machine learning algorithms in a multi-modal style. The domain is fast gaining traction owing to its applications across multiple disciplines with crucial gains. Research limitations/implications The research is focused on deep learning and computer vision-based techniques to generate a context for assistive driving and it would definitely adopt by the ADAS manufacturers. Social implications As context-aware assistive driving would work in real-time and it would save the lives of many drivers, pedestrians. Originality/value This paper provides an understanding of context-aware deep learning frameworks for assistive driving. The research is mainly focused on deep learning and computer vision-based techniques to generate a context for assistive driving. It incorporates the latest state-of-the-art techniques using suitable driving context and the driver is alerted. Many automobile manufacturing companies and researchers would refer to this study for their enhancements.

In last few decades, technological revolution has accelerated the deployment of large scale surveillance systems on almost all public places such as malls, hospitals, airports, railways, bus stations, roads, etc. These intelligent surveillance systems can play crucial role in governance of situations, collective security and safety, mitigating as well as prevention of adversaries. With gradual increase in multi camera surveillance systems enclosing multi angle views of same as well as different scenes has increased complexity of monitoring the systems by manual inspection. Abnormalities also known as anomalies or outliers are inevitable part of the existence and presumed to be rare in occurrence. Manual monitoring of such abnormalities is susceptible to errors and limited by human capabilities such as inattention and tiresome. Hence in the field of computer vision, automated abnormal activity recognition (AAR) from surveillance systems is emerging research area. The intent of this research is to shed a light on recent innovations and developments that have made a mark in abnormal activity recognition (AAR) involving deep learning. This paper also includes conventional categorization of anomalies based on different perspectives which can provide better understanding to young researchers. Though recent developments still poses many real time challenges in automatic abnormal activity recognition, some of them are enclosed in this paper.

AbstractGathering a large number of people in a shared physical area is very common in urban culture. Although there are limitless examples of mega crowds, the Islamic religious ritual, the Hajj, is considered as one of the greatest crowd scenarios in the world. The Hajj is carried out once in a year with a congregation of millions of people when the Muslims visit the holy city of Makkah at a given time and date. Such a big crowd is always prone to public safety issues, and therefore requires proper measures to ensure safe and comfortable arrangement. Through the advances in computer vision based scene understanding, automatic analysis of crowd scenes is gaining popularity. However, existing crowd analysis algorithms might not be able to correctly interpret the video content in the context of the Hajj. This is because the Hajj is a unique congregation of millions of people crowded in a small area, which can overwhelm the use of existing video and computer vision based sophisticated algorithms. Through our studies on crowd analysis, crowd counting, density estimation, and the Hajj crowd behavior, we faced the need of a review work to get a research direction for abnormal behavior analysis of Hajj pilgrims. Therefore, this review aims to summarize the research works relevant to the broader field of video analytics using deep learning with a special focus on the visual surveillance in the Hajj. The review identifies the challenges and leading-edge techniques of visual surveillance in general, which may gracefully be adaptable to the applications of Hajj and Umrah. The paper presents detailed reviews on existing techniques and approaches employed for crowd analysis from crowd videos, specifically the techniques that use deep learning in detecting abnormal behavior. These observations give us the impetus to undertake a painstaking yet exhilarating journey on crowd analysis, classification and detection of any abnormal movement of the Hajj pilgrims. Furthermore, because the Hajj pilgrimage is the most crowded domain for video-related extensive research activities, this study motivates us to critically analyze the crowd on a large scale.

Anomaly detection is a branch of behavior understanding in surveillance scenes, where anomalies represent a deviation in the behavior of scene entities (viz.,humans, vehicles, and environment) from regular patterns. In pedestrian walkways, this plays a vital role in enhancing safety. With the widespread use of video surveillance systems and the escalating video volume, manual examination of abnormal events becomes time-intensive.Hence, the need for an automated surveillance system adept at anomaly detection is crucial, especially within the realm of computer vision (CV) research. The surge in interest towards deep learning (DL) algorithms has significantly impacted CV techniques, including object detection and classification. Unlike traditional reliance on supervised learning requiring labeled datasets, DL offers advancements in these applications. Thus, this study presents an Optimal Deep Transfer Learning Enabled Object Detector for Anomaly Recognition in Pedestrian Ways (ODTLOD-ARPW) technique. The purpose of the ODTLOD-ARPW method is to recognize the occurrence of anomalies in pedestrian walkways using a DL-based object detector. In the ODTLOD-ARPW technique, the image pre-processing initially takes place using two sub-processes namely Wiener filtering (WF) based pre-processing and dynamic histogram equalization-based contrast enhancement. For anomaly detection, the ODTLOD-ARPW technique employs the YOLOV8s model which offers enhanced accuracy and performance. The hyperparameter tuning process takes place using a root mean square propagation (RMSProp) optimizer. The performance analysis of the ODTLOD-ARPW method is tested under the UCSD anomaly detection dataset. An extensive comparative study reported that the ODTLOD-ARPW technique reaches an effective performance with other models with maximum accuracy of 98.67%.

Traffic signs are a mandatory feature of road traffic regulations worldwide. They are responsible for slowing down the speed and carrying out many other valuable duties, notifying drivers about dangerous parts of the route, signaling traffic destination, prohibiting, or allowing passage. In this way, traffic is smoother, it becomes better regulated and drivers understand, mark, and interpret the rules well. For this purpose, Machine Learning (ML) study is carried out with the Deep Learning (DL) approach, the Real-Time (RT) Traffic Signs Recognition (TSR) is successfully developed, and a 99,68% test accuracy is obtained which has been gradually built on autonomous vehicles. It is developed to alert the driver for traffic signs appearing on the road and it is assisted the driver reach the speed limit set in the section of the lane, ride, overtaking, etc. The developed TSR helps to boost safety dramatically on the way to autonomous driving. The system is built on the DL, and it is trained with the Convolutional Neural Network (CNN) model to a classifier and predicts the status which is very effective for image classification purposes, and it is the most common and lovable algorithm for image data processing. It is also visualized how the accuracy and loss rates have changed over time. Later, it is implemented the graphical Unit Interface (GUI) were to show the results and draw the accuracy and the loss graphs. To realize classification as RT, Computer Vision (CV) approach is also included within the developed software to support camera viewing and understanding digital images.

Multiple Vehicle detection and tracking is one of the hot research topics in the field of intelligent transportation systems, image processing, computer vision, robotics whereas applications are real time traffic monitoring, lane estimation, accident avoidance, alarm signal to indicate road accidents to save the public safety and so on. There exists a numerous higher level applications are motivated by a young researchers and scientists to identify the newly advanced techniques in which to solve the real time traffic problems using machine learning and deep learning methods to track multiple vehicles accurately. To addresses the various existing challenges in machine learning and deep learning based multiple vehicle detection and tracking algorithms namely camera oscillation, shadowing, changing in background motion, cluttering, camouflage etc. for the detection rate decreases dramatically when the distribution of the training samples and the scene target samples do not match. To address this issue, a new hybrid model of two-tier classifier of Haar+HOG, SVM+AdaBoost classifier algorithm based on a feature extraction algorithm is proposed in this paper. Inspired by the Adaptive Discrete Classifiers mechanism multiple relatively independent source samples are first used to build multiple classifiers and then particle grouping is used to generate the target training samples with confident scores. The global manual feature extraction ability of deep convolutional neural network is then used to perform source-target scene feature similarity calculation with a deep auto encoder in order to design a composite deep structure based adaptive discrete classifier and its global training method. The main contributions of this paper are threefold: 1) To improve the overall accuracy rate of multiple vehicle detection and tracking of front-view vehicles alone rather than full-sided vehicles. 2) The novelty of our proposed work is for particle grouping of multi-vehicles such as car, bus and lorry. 3) To propose the tracking of front- view multi- vehicles in linear and non-linear motion using particle and extended kalman filter along with hybrid new multi-vehicle tracking algorithm and attains 93.6% of accuracy is shown in the experimental results. We evaluates our proposed method with standard data sets PETS 2016 and 5 self-data sets iROAD were manually collected on traffic road and compared with the existing state of the art approaches and along with the Experiments on the Kitti dataset and a 3 different self -data set captured by our group demonstrate that the proposed method performs better than the existing machine-learning based vehicle detection methods. In addition, compared with the existing automatic feature extraction and region based object detection methods, our new hybrid method improves the overall detection rate by an average of approximately 5% of existing methods.

The use of hyperspectral imaging (HSI) for autonomous navigation is a promising field of research that aims to improve the accuracy and robustness of detection, tracking, and scene understanding systems based on vision sensors. The combination of advanced computer algorithms, such as deep neural networks (DNNs), and small-size snapshot HSI cameras allows to strengthen the reliability of those vision systems. Using HSI, some intrinsic limitations of greyscale and RGB imaging in depicting physical properties of targets related to the spectral reflectance of materials (metamerism) are overcome. Despite the promising results of many published HSI-based computer vision developments, the strict requirements of safety-critical applications such as autonomous driving systems (ADS) regarding latency, resource consumption, and security are prompting the migration of machine learning (ML)-based solutions to edge platforms. This involves a thorough software/hardware co-design scheme to distribute and optimize the tasks efficiently among the limited resources of computing platforms. With respect to inference, the over-parameterized nature of DNNs poses significant computational challenges for real-time on-the-edge deployment. In addition, the intensive data preprocessing required by HSI, which is frequently overlooked, must be carefully managed in terms of memory arrangement and inter-task communication to enable an efficient integrated pipeline design on a system on chip (SoC). This work presents a set of optimization techniques for the practical co-design of a DNN-based HSI segmentation processor deployed on a field programmable gate array (FPGA)-based SoC targeted at ADS, including key optimizations such as functional software/hardware task distribution, hardware-aware preprocessing, ML model compression, and a complete pipelined deployment. Applied compression techniques significantly reduce the complexity of the designed DNN to 24.34% of the original operations and to 1.02% of the original number of parameters, achieving a 2.86x speed-up in the inference task without noticeable degradation of the segmentation accuracy.

Susheel Kumar Sharma’s Unwinding Self: A Collection of Poems. Cuttack: Vishvanatha Kaviraj Institute, 2020, ISBN: 978-81-943450-3-9, Paperback, pp. viii + 152.
 Like his earlier collection, The Door is Half Open, Susheel Kumar Sharma’s Unwinding Self: A Collection of Poems has three sections consisting of forty-two poems of varied length and style, a detailed Glossary mainly on the proper nouns from Indian culture and tradition and seven Afterwords from the pens of the trained readers from different countries of four continents. The structure of the book is circular. The first poem “Snapshots” indicates fifteen kaleidoscopic patterns of different moods of life in about fifteen words each. It seems to be a rumination on the variegated images of everyday experiences ranging from individual concerns to spiritual values. Art-wise, they can be called mini-micro-poems as is the last poem of the book. 
 While the character limit in a micro poem is generally 140 (the character limit on Twitter) Susheel has used just around 65 in each of these poems. Naturally, imagery, symbolism and cinematic technique play a great role in this case. In “The End of the Road” the poet depicts his individual experiences particularly changing scenario of the world. He seems to be worried about his eyesight getting weak with the passage of time, simultaneously he contrasts the weakness of his eyesight with the hypocrisy permeating the human life. He compares his diminishing eyesight to Milton and shows his fear as if he will get blind. He changes his spectacles six times to clear his vision and see the plurality of a reality in human life. It is an irony on the changing aspects of human life causing miseries to the humanity. At the end of the poem, the poet admits the huge changes based on the sham principles: “The world has lost its original colour” (4). The concluding lines of the poem make a mockery of the people who are not able to recognise reality in the right perspective.
 The poem “Durga Puja in 2013” deals with the celebration of the festival “Durga Puja” popular in the Hindu religion. The poet’s urge to be with Ma Durga shows his dedication towards the Goddess Durga, whom he addresses with different names like ‘Mai’, ‘Ma’ and ‘Mother’. He worships her power and expresses deep reverence for annihilating the evil-spirits. The festival Durga Puja also reminds people of victory of the goddess on the elusive demons in the battlefield. “Chasing a Dream on the Ganges” is another poem having spiritual overtones. Similarly, the poem “Akshya Tritya” has religious and spiritual connotations. It reflects curiosity of people for celebration of “Akshya Tritya” with enthusiasm. But the political and economic overtones cannot be ignored as the poem ends with the remarkable comments:
 The GDP may go up on this day;
 Even, Budia is able to
 Eat to his fill; Panditji can blow his
 Conch shell with full might.
 Outside, somebody is asking for votes;
 Somebody is urging others to vote.
 I shall vote for Akshya Tritya. (65-66)
 “On Reading Langston Hughes’ ‘Theme for English B’” is a long poem in the collection. In this poem, the poet reveals a learner’s craving for learning, perhaps who comes from an extremely poor background to pursue his dreams of higher education. The poet considers the learner’s plights of early childhood, school education and evolutionary spirit. He associates it with Dronacharya and Eklavya to describe the mythical system of education. He does not want to be burdened with the self-guilt by denying the student to be his ‘guru’ therefore, he accepts the challenge to change his life. Finally, he shows his sympathy towards the learner and decides to be the ‘guru’: “It is better to face/A challenge and change/Than to be burden with a life/Of self-guilt. /I put my signatures on his form willy-nilly” (11).
 The poem “The Destitute” is an ironical presentation of the modern ways of living seeking pleasure in the exotic locations all over the world. It portrays the life of a person who has to leave his motherland for earning his livelihood, and has to face an irreparable loss affecting moral virtues, lifestyle, health and sometimes resulting in deaths. The poem “The Black Experience” deals with the suppression of the Africans by the white people. The poem “Me, A Black Doxy”, perhaps points out the dilemma of a black woman whether she should prostitute herself or not, to earn her livelihood. Perhaps, her deep consciousness about her self-esteem does not allow her to indulge in it but she thinks that she is not alone in objectifying herself for money in the street. Her voice resonates repeatedly with the guilt of her indulgence on the filthy streets:
 At the dining time
 Me not alone? In the crowded street
 Me not alone?
 They ’ave white, grey, pink hair
 Me ’ave black hair – me not alone
 There’s a crowd with black hair.
 Me ’ave no black money
 Me not alone? (14)
 The poem “Thus Spake a Woman” is structured in five sections having expressions of the different aspects of a woman’s love designs. It depicts a woman’s dreams and her attraction towards her lover. The auditory images like “strings of a violin”, “music of the violin” and “clinch in my fist” multiply intensity of her feelings. With development of the poem, her dreams seem to be shattered and sadness know the doors of her dreamland. Finally, she is confronted with sadness and is taken back to the past memories reminding her of the difficult situations she had faced.
 Replete with poetic irony, “Bubli Poems” presents the journey of a female, who, from the formative years of her life to womanhood, experienced gender stereotypes, biased sociocultural practices, and ephemeral happiness on the faces of other girls around her. The poem showcases the transformation of a village girl into a New Woman, who dreams her existence in all types of luxurious belongings rather than identifying her independent existence and finding out her own ways of living. Her dreams lead her to social mobility through education, friendships, and the freedom that she gains from her parents, family, society and culture. She attempts her luck in the different walks of human life, particularly singing and dancing and imagines her social status and wide popularity similar to those of the famous Indian actresses viz. Katrina and Madhuri Dixit: “One day Bubli was standing before the mirror/Putting on a jeans and jacket and shaking her hips/She was trying to be a local Katrina” (41). She readily bears the freakish behaviour of the rustic/uncultured lads, derogatory comments, and physical assaults in order to fulfil her expectations and achieves her individual freedom. Having enjoyed all the worldly happiness and fashionable life, ultimately, she is confronted with the evils designs around her which make her worried, as if she is ignorant of the world replete with the evils and agonies: “Bubli was ignorant of her agony and the lost calm” (42). The examples of direct poetic irony and ironic expressions of the socio-cultural evils, and the different governing bodies globally, are explicit in this poem: “Bubli is a leader/What though if a cheerleader./The news makes her family happy.”(40), “Others were blaming the Vice-Chancellor/ Some others the system;/ Some the freedom given to girls;”(45), and “Some blame poverty; some the IMF;/ Some the UN; some the environment;/ Some the arms race; some the crony’s lust;/ Some the US’s craving for power;/Some the UK’s greed. (46-47). Finally, Bubli finds that her imaginative world is fragile. She gives up her corporeal dreams which have taken the peace of her mind away. She yearns for shelter in the temples and churches and surrenders herself before deities praying for her liberation: “Jai Kali,/ Jai Mahakali, Jai Ma, Jai Jagaddhatri,/ Save me, save the world.” (47).
 In the poem “The Unlucky”, the poet jibes at those who are lethargic in reading. He identifies four kinds of readers and places himself in the fourth category by rating himself a ‘poor’ reader. The first three categories remind the readers of William Shakespeare’s statement “Some are born great, some achieve greatness, and some have greatness thrust upon them.” At the end of the poem, the poet questions himself for being a poet and teacher. The question itself reflects on his ironic presentation of himself as a poor reader because a poet’s wisdom is compared with that of the philosopher and everybody worships and bows before a teacher, a “guru”, in the Indian tradition. The poet is considered the embodiment of both. The poet’s unfulfilled wish to have been born in Prayagraj is indexed with compunction when the poem ends with the question “Why was I not born in Prayagraj?” (52). Ending with a question mark, the last line of the poem expresses his desire for perfection. The next poem, “Saying Goodbye”, is elegiac in tone and has an allusion to Thomas Gray’s “The Elegy Written in a Country Churchyard” in the line “When the curfew tolls the knell of the parting day”; it ends with a question mark. The poem seems to be a depiction of the essence and immortality of ‘time’. Reflecting on the poet’s consideration of the power and beauty of ‘time’, Pradeep Kumar Patra rightly points out, “It is such a phenomena that nobody can turn away from it. The moment is both beautiful as well as ferocious. It beautifies and showcases everything and at the same time pulls everything down when necessary” (146). Apparently, the poem “The Kerala Flood 2018”is an expression of emotions at the disaster caused by the flood in 2018. By reminding of Gandhi’s tenets to be followed by people for the sake of morality and humankind, the poet makes an implicit criticism of the pretentions, and violation of pledges made by people to care of other beings, particularly, cow that is worshiped as “mother” and is considered to be a symbol of fertility, peace and holiness in Hinduism as well as the Buddhist culture. The poet also denigrates people who deliberately ignore the sanctity of the human life in Hinduism and slaughter the animal cow to satisfy their appetites. In the poem, the carnivorous are criticized explicitly, but those who pretend to be herbivorous are decried as shams:
 If a cow is sacrosanct
 And people eat beef
 One has to take a side.
 Some of the friends chose to
 Side with cow and others
 With the beef-eaters.
 Some were more human
 They chose both. (55)
 The poet infuses positivity into the minds of the Indian people. Perhaps, he thinks that, for Indians, poverty, ignorance, dirt and mud are not taboos as if they are habitual to forbear evils by their instincts. They readily accept them and live their lives happily with pride considering their deity as the preserver of their lives. The poem “A Family by the Road” is an example of such beliefs, in which the poet lavishes most of his poetic depiction on the significance of the Lord Shiva, the preserver of people in Hinduism:
 Let me enjoy my freedom.
 I am proud of my poverty.
 I am proud of my ignorance.
 I am proud of my dirt.
 I have a home because of these.
 I am proud of my home.
 My future is writ on the walls
 Of your houses
 My family shall stay in the mud.
 After all, somebody is needed
 To clean the dirt as well.
 I am Shiva,
 Shivoham. (73)
 In the poem “Kabir’s Chadar”, the poet invokes several virtues to back up his faith in spirituality and simplicity. He draws a line of merit and virtue between Kabir’s Chadar which is ‘white’ and his own which is “thickly woven” and “Patterned with various beautiful designs/ In dark but shining colours” (50). The poet expresses his views on Kabir’s ‘white’ Chadar symbolically to inculcate the sense of purity, fortitude, spirituality, and righteousness among people. The purpose of his direct comparison between them is to refute artificiality, guilt and evil intents of humanity, and propagate spiritual purity, the stark simplicities of our old way of life, and follow the patience of a saint like Kabir.
 The poem “Distancing” is a statement of poetic irony on the city having two different names known as Bombay and Mumbai. The poet sneers at its existence in Atlas. Although the poet portraits the historical events jeering at the distancing between the two cities as if they are really different, yet the poet’s prophetic anticipation about the spread of the COVID-19 in India cannot be denied prima facie. The poet’s overwhelming opinions on the overcrowded city of Bombay warn humankind to rescue their lives. Even though the poem seems to have individual expressions of the poet, leaves a message of distancing to be understood by the people for their safety against the uneven things. The poem “Crowded Locals” seems to be a sequel to the poem “Distancing”. Although the poet’s purpose, and appeal to the commonplace for distancing cannot be affirmed by the readers yet his remarks on the overcrowded cities like in Mumbai (“Crowded Locals”), foresee some risk to the humankind. In the poem “Crowded Locals”, he details the mobility of people from one place to another, having dreams in their eyes and puzzles in their minds for their livelihood while feeling insecure especially, pickpockets, thieves and strangers. The poet also makes sneering comments on the body odour of people travelling in first class. However, these two poems have become a novel contribution for social distancing to fight against the COVID-19.
 In the poem “Buy Books, Not Diamonds” the poet makes an ironical interpretation of social anarchy, political upheaval, and threat of violence. In this poem, the poet vies attention of the readers towards the socio-cultural anarchy, especially, anarchy falls on the academic institutions in the western countries where capitalism, aristocracy, dictatorship have armed children not with books which inculcate human values but with rifles which create fear and cause violence resulting in deaths. The poet’s perplexed opinions find manifestation in such a way as if books have been replaced with diamonds and guns, therefore, human values are on the verge of collapse: “Nine radiant diamonds are no match/ To the redness of the queen of spades. . . . / … holding/ Rifles is a better option than/ Hawking groundnuts on the streets?” (67).The poet also decries the spread of austere religious practices and jihadist movement like Boko Haram, powerful personalities, regulatory bodies and religious persons: “Boko Haram has come/Obama has also come/The UN has come/Even John has come with/Various kinds of ointments” (67).
 The poem “Lost Childhood” seems to be a memoir in which the poet compares the early life of an orphan with the child who enjoys early years of their lives under the safety of their parents. Similarly, the theme of the poem “Hands” deals with the poet’s past experiences of the lifestyle and its comparison to the present generation. The poet’s deep reverence for his parents reveals his clear understanding of the ways of living and human values. He seems to be very grateful to his father as if he wants to make his life peaceful by reading the lines of his palms: “I need to read the lines in his palm” (70).
 In the poem “A Gush of Wind”, the poet deliberates on the role of Nature in our lives. The poem is divided into three sections, perhaps developing in three different forms of the wind viz. air, storm, and breeze respectively. It is structured around the significance of the Nature. In the first section, the poet lays emphasis on the air we breathe and keep ourselves fresh as if it is a panacea. The poet criticizes artificial and material things like AC. In the second section, he depicts the stormy nature of the wind scattering papers, making the bed sheets dusty affecting or breaking the different types of fragile and luxurious objects like Italian carpets and lamp shades with its strong blow entering the oriels and window panes of the houses. Apparently, the poem may be an individual expression, but it seems to be a caricature on the majesty of the rich people who ignore the use of eco-chic objects and disobey the Nature’s behest. In the third and the last section of the poem, the poet’s tone is critical towards Whitman, Pushkin and Ginsberg for their pseudoscientific philosophy of adherence to the Nature. Finally, he opens himself to enjoy the wind fearlessly.
 The poems like “A Voice” , “The New Year Dawn”, “The New Age”, “The World in Words in 2015”, “A Pond Nearby”, “Wearing the Scarlet Letter ‘A’”, “A Mock Drill”, “Strutting Around”, “Sahibs, Snobs, Sinners”, “Endless Wait”, “The Soul with a New Hat”, “Renewed Hope”, “Like Father, Unlike Son”, “Hands”, “Rechristening the City”, “Coffee”, “The Unborn Poem”, “The Fountain Square”, “Ram Setu”, and “Connaught Place” touch upon the different themes. These poems reveal poet’s creativity and unique features of his poetic arts and crafts.
 The last poem of the collection “Stories from the Mahabharata” is written in twenty-five stanzas consisting of three lines each. Each stanza either describes a scene or narrates a story from the Mahabharata, the source of the poem. Every stanza has an independent action verb to describe the actions of different characters drawn from the Mahabharata. Thus, each stanza is a complete miniscule poem in itself which seems to be a remarkable characteristic of the poem. It is an exquisite example of ‘Micro-poetry’ on paper, remarkable for its brevity, dexterity and intensity. The poet’s conscious and brilliant reframing of the stories in his poem sets an example of a new type of ‘Found Poetry’ for his readers.
 Although the poet’s use of various types images—natural, comic, tragic, childhood, horticultural, retains the attention of readers yet the abundant evidences of anaphora reflect redundancy and affect the readers’ concentration and diminishes their mental perception, for examples, pronouns ‘her’ and ‘we’ in a very small poem “Lost Childhood”, articles ‘the’ and ‘all’ in “Crowded Locals”, the phrase ‘I am proud of’ in “A Family by the Road” occur many times. Svitlana Buchatska’s concise but evaluative views in her Afterword to Unwinding Self help the readers to catch hold of the poet’s depiction of his emotions. She writes, “Being a keen observer of life he vividly depicts people’s life, traditions and emotions involving us into their rich spiritual world. His poems are the reflection on the Master’s world of values, love to his family, friends, students and what is more, to his beloved India. Thus, the author reveals all his beliefs, attitudes, myths and allusions which are the patterns used by the Indian poets” (150).
 W. H. Auden defines poetry as “the clear expression of mixed feelings.” It seems so true of Susheel Sharma’s Unwinding Self. It is a mixture of poems that touch upon the different aspects of human life. It can be averred that the collection consists of the poet’s seamless efforts to delve into the various domains of the human life and spot for the different places as well. It is a poetic revue in verse in which the poet instils energy, confidence, power and enthusiasm into minds of Indian people and touches upon all aspects of their lives. The poverty, ignorance, dirt, mud, daily struggle against liars, thieves, pickpockets, touts, politician and darkness have been depicted not as weaknesses of people in Indian culture but their strengths, because they have courage to overcome darkness and see the advent of a new era. The poems teach people morality, guide them to relive their pains and lead them to their salvation. Patricia Prime’s opinion is remarkable: “Sharma writes about his family, men and women, childhood, identity, roots and rootlessness, memory and loss, dreams and interactions with nature and place. His poised, articulate poems are remarkable for their wit, conversational tone and insight” (138). Through the poems in the collection, the poet dovetails the niceties of the Indian culture, and communicates its beauty and uniqueness meticulously. The language of the poem is lucid, elevated and eloquent. The poet’s use of diction seems to be very simple and colloquial like that of an inspiring teacher.
 On the whole the book is more than just a collection of poems as it teaches the readers a lot about the world around them through a detailed Glossary appended soon after the poems in the collection. It provides supplementary information about the terms used abundantly in Indian scriptures, myths, and other religious and academic writings. The Glossary, therefore, plays pivotal role in unfolding the layers of meaning and reaching the hearts of the global readers. The “Afterwords” appended at the end, enhances readability of poems and displays worldwide acceptability, intelligibility, and popularity of the poet. The Afterwords are a good example of authentic Formalistic criticism and New Criticism. They indirectly teach a formative reader and critic the importance of forming one’s opinion, direct reading and writing without any crutches of the critics.