Search

The oil palm plantation industry heavily relies on foreign labour for harvesting, particularly for collecting loose fruits (LF) alongside fresh fruit bunches (FFB). Manual LF collection, involving bending and repetitive movements, not only diminishes productivity but also poses health risks to workers. This study proposes an automated LF collector to mitigate these challenges. The developed system integrates an LF picker with a robot arm, an LF detector using image processing and a camera, a GPS-based human-follower vehicle, a back-to-home navigation system based on weight detection, and an obstacle avoidance system. The automated LF collector aims to operate autonomously, reducing workforce reliance and enhancing productivity in oil palm plantations. The study discusses the motivation, challenges, and objectives of developing such a system, emphasizing its potential economic and societal benefits. Additionally, the implementation of a novel image processing technique, Faster Objects More Objects (FOMO), using a neural network, is detailed for efficient LF detection. The proposed automated LF collector addresses labour shortages, enhances economic productivity, and reduces the risk of worker injuries in the oil palm industry. [ABSTRACT FROM AUTHOR]

Driver drowsiness and fatigue are leading causes of road accidents, particularly due to drivers having to operate vehicles for long hours under challenging physical and adverse weather conditions. Despite extensive research on subjective, physiological, and vehicle-based detection methods, recent reports still highlight drowsiness and fatigue as significant contributors to accidents and injuries on highways. However, existing drowsiness detection models are often intrusive or expensive. To address these limitations, we propose a low-cost, non-intrusive drowsiness detection model based on facial features, leveraging principles of image processing and machine learning. Our research involves a multi-step approach. Initially, image frames are extracted and subjected to image processing techniques. The images are subsequently passed through a Dlib classifier to detect and localize 68 facial landmarks, including the mouth and eyes. From the localized eye landmarks, salient features such as eye aspect ratio (EAR), pupil circularity (PUC), and eyelid closure over the eye pupil are extracted. To classify the driver's state, these collective features are inputted into various machine learning classifiers. We conducted experiments using the National Tsing Hua University Computer Vision Lab dataset, implementing the proposed approach in the Python OpenCV environment. Notably, the KNN algorithm demonstrated the highest accuracy when the test data was evaluated, yielding a paramount accuracy of 78% on the NTHU DDD dataset. In conclusion, the non-intrusive facial features-based driver drowsiness detection model offers a cost-effective solution to address the pervasive problem of drowsiness-related accidents. The promising results obtained through the utilization of machine learning techniques underscore the potential for practical implementation. [ABSTRACT FROM AUTHOR]

There is a correlation between our posture and our physical and emotional wellness. The detection of various human postures has been approached in a variety of ways. Determine a patient's resting position, for example, using posture analysis in the medical industry. Image processing using OpenCV and mediapipe python library for human posture estimation. Analyzing standing and sitting postures is made easier with an image-processing-based technique. The benefits of fitness activities to a person's health are remarkable, but if they are done improperly, they may be ineffective and even detrimental. Exercise blunders happen when someone doesn't adopt the right posture. This suggested application makes use of pose estimation to identify the user's workout posture and then offers specific, tailored advice on what the user can do to correct their posture. A pose estimation module in python called mediapipe helps to detect all major joints of the human body. Then it calculates the angle between the joints and increases the counter for each repetition. Pose estimate uses a picture or video of the subject to determine the precise locations of the body's major joints. This computer vision technology recognizes human posture in films and displays important regions, such as the elbow or shoulder, in the finished product. Another way to work out at home is through interactive games. The proposed application has an interactive boxing game built using pygame, which helps all age groups to work out and take care of their health. Overall the user can track all their details on a workout on this application, like how many calories are burned and the daily goals, and different workout planners for each individual. [ABSTRACT FROM AUTHOR]

Among all foods, rice is the one that people from different cultures eat the most of. When rice is of high grade, its price rises sharply. As it is, a physical evaluation method using the unaided eye is used to evaluate the rice variety & grade. Nevertheless, this approach is laborious, time-consuming, depends on human expertise, and poses risks to the investigator's wellbeing. This paper proposes a technique that employs computerized methods of image processing to automatically detect and categorise rice grains, therefore addressing the aforementioned problems. This image processing method is ideal because it does not involve any physical touch while photographing the rice grains. To evaluate rice quality, a CNN is utilised to pre-process, partition, & retrieve characteristics from the capturand SVM algorithmic classifiers. Our comparative analysis shows that the suggested system categorization outperforms the alternative. [ABSTRACT FROM AUTHOR]

Effective communication is crucial in our daily lives, and it occurs through various channels such as vocal, written, and body language. However, individuals with hearing impairments often rely on sign language as the primary means of communication. The inability to understand sign language can lead to isolation and barriers in communication, hindering the social lives of deaf individuals. To address this need, we propose a marker-free, visual Indian Sign Language identification system that employs image processing, computer vision, and neural network techniques. Our proposed system analyzes video footage captured by a webcam to recognize hand gestures and translate them into text, which is subsequently converted into audio. The system uses a range of image processing techniques to identify the shape of the hand from continuous video frames, including background subtraction, thresholding, and contour detection. The Haar Cascade Classifier algorithm is used to interpret the signs and assign meaning to them based on the recognized patterns. Finally, a speech synthesizer is employed to convert the displayed text into speech. The proposed system is intended to improve the social lives of deaf individuals by facilitating communication with hearing individuals. It is designed to be user-friendly, efficient, and affordable, as it does not require any additional hardware or markers to recognize signs. The proposed system could be integrated into various devices such as smartphones, tablets, or laptops, making it accessible to a wide range of users. The implementation of such a system could potentially break down communication barriers between the deaf and hearing communities, providing deaf individuals with more opportunities to interact with others and participate in society. [ABSTRACT FROM AUTHOR]

As the number of vehicles increases day by day, finding an empty parking space to park a vehicle becomes difficult. Car parking can cause wasted time and interfere with surrounding mobility, as a result the parking area cannot be utilized optimally. The existing parking management system using sensors to detect the available parking spaces is less effective and efficient, for example a system with the use of ultrasonic sensors that must be placed in each parking box will require many sensors in large-scale implementation. The proposed intelligent parking system provides a structured solution by using a parking lot camera available in the campus or office area to observe the used parking area and using image processing to detect the available parking space from the camera in real time. From the results of image processing available parking spaces will be recommended to users through a front-end system based on the closest distance that helps drivers park their vehicles. The proposed system improves the overall effectiveness and efficiency of the current parking system and solves the problem that drivers spend a lot of time in finding suitable parking spaces in crowded campus or office parking areas. The architecture of the intelligent parking system includes three stages: the first stage of the system uses sensors to capture images of the parking area and sends them to the database server in real time; the second stage of the proposed method uses object detection algorithms (ie, Region-based Convolutional Neural Networks) to identify whether parking spaces in the building area are available or not and calculate their utility; the third stage a front-end system was developed for drivers to get real-time parking information by using a monitor placed at the entrance gate of a campus or office parking area. [ABSTRACT FROM AUTHOR]

Currently, disease recognition is mostly based on the use of modern technology. Image processing system is one of the technologies that can be used in the process of detecting various diseases. The aim of this research is to build a system that can be used to detect COVID-19 infections automatically using image processing. The Covid-19 disease in the human body has a specific area in the body, namely the lungs. Modern medical science is currently so advanced that X-ray machine can provide digital images of the lungs and with the help of image processing, these images can be identified as normal or infected with COVID-19. The classification method used is logistic regression using parameters of contrast, correlation, energy, homogeneity and entropy. The results show that the system built has succeeded in detecting normal lungs and lungs infected with Covid-19 with 94 percent accuracy [ABSTRACT FROM AUTHOR]

The development of world technology is growing rapidly, including the health sector in the form of a tool for detecting various kinds of diseases. These technologies include systems that can recognize images and perform classification automatically. An automated system like this is called an intelligent system which is very useful in the health sector because it can help even replace a doctor or health expert in carrying out the task of diagnosing disease. The aim of this research is to build a system that can be used to detect COVID-19 infections automatically using image processing. The classification method used is an image processing technique using machine learning, namely a support vector machine (SVM) with entropy and correlation parameters. The image that will be recognized is the image of the lungs from the X-ray machine. The results show that the system built has succeeded in detecting normal lungs and lungs infected with Covid-19 with 98 percent accuracy [ABSTRACT FROM AUTHOR]

In order to maintain good health and avoid illness, people need access to potable water. People often become sick from water-related causes because they drink polluted water. This is why monitoring the water quality is essential. The Novel Tensor model is built using time-, location-, and water-related variables. A support vector regression approach is used to forecast the novel water quality by the suggested system, which takes water samples from the DWS system. This research employed a dataset consisting of about 1993 samples, with 127 samples analyzed statistically using SPSS. The results suggest that this new tensor has the potential to produce the best results with the most accuracy (p=.676). The goal of this research is to improve water quality monitoring by using a novel approach to existing image processing classifiers. When compared to older, less sophisticated algorithms, modern artificial neural network methods perform far better. [ABSTRACT FROM AUTHOR]

Counterfeiting of quality rice was rife in Indonesia. This research was conducted to develop technology to identify differences in premium and non-premium rice quality based on pre-existing digital images. Artificial neural networks and digital image processing methods to identify premium and medium (non-premium) rice quality were applied in this research. Statistical analysis of this study used the SPSS program. This research is observation-type research. This research design uses an artificial neural network with uses 3 layers, namely the results of shape feature extraction on the metric, eccentricity, area, and perimeter parameters as input or input layers, hidden or hidden layers, and premium rice and non-premium (medium) rice as output or output layers. This research uses 52 images as training and 20 images as testing. The obtained image was taken at a distance of 25 cm. This research showed that the results of training using artificial neural networks (ANN) on 52 images obtained an accuracy of 92%. The test results using 20 images obtained 95% accuracy, 63.33% sensitivity, and 10% specificity. Based on statistical analysis using the Mann-Whitney test, it obtained the asymph value. Sig (2-tailed) < 0.05 indicates the difference between premium and non-premium rice using metric, eccentricity, perimeter, and area parameters. [ABSTRACT FROM AUTHOR]

Applications like underwater exploration, marine life monitoring, and underwater item detection depend heavily on the ability to segment underwater images. However, underwater habitats' varied sights and challenging qualities, such as reduced visibility, make proper segmentation challenging. In this study, the performance of an encoder- decoder architecture, especially the Unet model, is compared to that of conventional image processing methods for binary segmentation of underwater images. Performance evaluation is based on critical criteria, such as sensitivity (recall), F1 score, precision, and accuracy. The evaluation's findings show that, among the methods examined, the Unet model has the highest sensitivity (95.16%). This demonstrates how well it works to locate bright areas in underwater photos. The Unet model also achieves a noteworthy F1 score of 86.71%, illustrating a favorable balance between precision and recall. While the Random Forest method's F1 score is slightly lower at 78.80%, it still displays comparable precision (75.46%) and accuracy (79.45%) values. The Otsu Level set technique performs worse on all metrics. The encoder-decoder architecture of the Unet model, which efficiently absorbs and uses both local and global contextual information throughout the segmentation process, can be credited with the model's better performance. [ABSTRACT FROM AUTHOR]

The photos taken by all, bring backs remarkable memories for everyone life. By keeping in mind, this research was focused to enhance the quality of an image, aiming to bring it as close to its original high-quality state. Advanced image processing techniques are employed to achieve this goal. This approach is particularly useful when dealing with degraded images that remain usable but require improvement to enhance their utility or presentation. To elevate the Naturalness Image Quality Evaluator (NIQE) score, we employ the Novel Generative Adversarial Network v3 (NGAN3) with Enhanced Super Resolution Generative Adversarial Networks (ESRGAN). Our study involved two distinct groups: Group 1, which utilized the NGAN3 algorithm, and Group 2, which employed the ESRGAN algorithm. The group sample size carefully determined using the Clincalc tool. This tool also facilitated the calculation of error rates, including an beta level of 0.2, alpha level of 0.05 with power of 0.05. In total, we analyzed 40 samples (20 per group). The NGAN3 algorithm yielded an average NIQE score of 4.46, while the ESRGAN algorithm produced an average NIQE score of 6.36. Notably, the ESRGAN algorithm demonstrated statistical significance with a significance value (P) of 0.001, as determined by a sample t-test. This result underscores the superiority of NGAN3 in terms of image quality. Furthermore, Novel GAN3 generates visually compelling and realistic images, surpassing the existing algorithm. Consequently, it holds the potential to enhance employment prospects by increasing the likelihood of securing well-paid jobs. [ABSTRACT FROM AUTHOR]

This research aims to explore the potential of various facial image types in diagnosing Autism Spectrum Disorder (ASD) through the application of deep learning neural networks. It delves into how deep learning algorithms perform with different facial image modalities, especially 2D and 3D, while addressing specific challenges associated with each. The methodology includes training deep learning models on distinct datasets and conducting an in-depth analysis of their accuracy and performance metrics. Significantly, the ResNet50V2 model recorded a 96.9% accuracy rate on the 2D dataset, and the Xception model achieved an 84.4% accuracy rate on the 3D dataset. These findings emphasize the strong capability of deep learning neural networks in making accurate ASD diagnoses from facial images. Nonetheless, the research reveals a stronger proficiency in handling 2D over 3D images, suggesting a need for more comprehensive 3D dataset training to improve three-dimensional image processing. Through evaluating the efficacy of different image modalities, this investigation enriches the field's knowledge base, highlights the necessity for robust dataset development, and charts a course for future studies to advance the precision and practicality of ASD diagnosis via deep learning approaches. [ABSTRACT FROM AUTHOR]

The oil palm plantation industry plays a pivotal role in Malaysia's economy, contributing significantly to global palm oil production and exports. However, the industry faces challenges related to labour-intensive harvesting, particularly the manual collection of loose oil palm fruits (LF). This manual process not only leads to reduced productivity but also poses health risks to workers. To address these issues, the development of an autonomous LF collector is proposed, aiming to reduce labour requirements and enhance collection efficiency. The automated LF collector incorporates a robotic arm for fruit picking, an LF detection system using image processing, a GPS-based human-follower vehicle, a back-to-home navigation system, and obstacle avoidance mechanisms. This innovation allows the LF collector to operate autonomously in oil palm plantations, thus relieving the workforce. The project details materials and components selection for the LF picker, providing insights into the practical implementation of this technology. By reducing the reliance on manual labour and improving LF collection, the proposed autonomous LF collector contributes to increasing overall palm oil production, ensuring the industry's sustainability and economic growth. [ABSTRACT FROM AUTHOR]

This work involves the development of a computer application known as the Emotion-Based Music Player, which is aimed specifically at music enthusiasts. Because choosing music can be difficult, most people prefer to play the playlist's tracks at random. Some of the songs selected as a result did not accurately reflect the users' current mood. Furthermore, no widely used music player can play songs based on the user's emotions. The proposed model can identify the user's emotions by extracting the user's facial expression. The songs in the suggested model will then be played by the music player based on the type of emotion identified. It aims to give music fans more ways to enjoy music. The proposed model encompasses a wide range of emotions, including Neutral, Sad, Happy, and Angry. The system heavily employs facial recognition and image processing technology using Fisher Face Method. The model's train and test inputs are still photographs in.jpeg format that is available online. To assess the proposed model's accuracy in identifying emotions, eighty users of various ages and genders—twenty for each type of emotion— are tested. According to the testing results, the proposed model has an 85% recognition rate. [ABSTRACT FROM AUTHOR]

Many devices, including digital cameras, washing machines, smart cars, computers, etc., will have a PCB as an integral component. Today's PCB demands the highest density of components and increased tolerance on these components. Surface mount technology (SMT) is making it possible to add many miniaturized components to PCBs. Component size is decreasing, and the component number is increasing; inspecting these components is becoming even more challenging. As many products are becoming smart, there is a huge demand for PCBs. Huge demands for PCBs also demand quicker inspection techniques. This shows the increased demand for online and inspection techniques. Many types of defects are encountered in the manufacturing of PCBs, such as missing components, cracks, etc. Many of the existing inspection techniques are not suitable for online PCB evaluation. In this scenario, the method presented in this article becomes very important. The present method uses image processing and deep learning-based technique for image identification and classification. Also, the method helps in the real-time assessment of PCBs. [ABSTRACT FROM AUTHOR]

Visual Surveillance is one of the necessary research fields in Computer vision and image processing. Various techniques are developed to perform automated video surveillance. One of the critical challenges in surveillance is detecting abandoned objects in crowded places like parks, shopping complexes and airports, etc., to prevent ecological loss frombomb blasting, gun fires, lost baggage, etc. Closed Circuit Television (CCTV) continuously monitor public places and discriminate unattended and abandoned luggage from different objects. An abandoned thing in surveillance is nothing but an unattended object without a person for a long duration. Identifying these objects in real-time can help prevent bomb blasts and terrorist attacks etc. Many types of research have been performed and published in surveillance for abandoned object detection in recent years. Deep learning is one of the significant fields where various techniques and algorithms are developed for detecting unattended objects. Though these methods detect the objects significantly, they lack computational complexity. A classical edge detection method using Canny edge detection to overcome these challenges is developed to identify the abandoned objects in surveillance scenes. [ABSTRACT FROM AUTHOR]

Deep learning's application to machine learning has seen significant growth in recent years. One of the greatest difficulties now confronting AI systems is the integration of visual data with machine learning in the field of medical diagnostics. Brain tumours, the result of unchecked cell proliferation, are a particularly lethal kind of illness. If painful brain tumours aren't treated properly, they may spread and cause other problems. Cancer treatment relies heavily on an accurate diagnosis. In order to determine whether a tumour is benign or malignant, accurate identification is required. More individuals are diagnosed with cancer now than ever before in part because physicians lack the knowledge to effectively treat tumours at an early stage. In this research, we classify brain MRI and PET scans into tumour and nontumor categories using a convolutional neural network (CNN) technique, augmented by Data Augmentation (DA) and Image Processing (IP) with Image Acquisition (IA). Although we only used a tiny dataset in our experiment, the high level of accuracy (96%) it achieved vouched for the efficiency and minimal complexity of our model. Our approach requires much less computational resources to achieve comparable levels of accuracy as compared to previous pre-trained models. [ABSTRACT FROM AUTHOR]

Tumours are currently the most common cancer cause. Many patients are at risk as a result of cancer. To detect tumours like brain tumours, the medical sector needs a quick, automated, effective, and trustworthy technique. A crucial part of treatment is detection. Doctors keep a patient out of danger if accurate tumour identification is possible. In this application, many image processing methods are applied. Doctors treat many tumour patients properly and save their lives using this application. Currently, medical professionals manually inspect the patient's Magnetic imaging scans of the brain to identify the position and magnitude of the brain neoplasm. This process is time-consuming and results in imprecise detection of the neoplasm. To identify brain tumours, we can apply a Deep Learning framework integrating Convolutional Neural Network (CNN), also called Neural Network, with Visual Geometry Group (VGG) 16 Transfer learning. The advantage of the model is to anticipate the existence of a tumour in an image by returning a positive result if present and a negative effect if not. [ABSTRACT FROM AUTHOR]

Recognizing objects is an important task in many real-world applications, including security systems, surveillance, and robots, making it a hot topic in computer vision research. Because of the progress made in deep learning, real-time object identification systems have been created. These systems can instantly identify things in photos and movies. A major focus of computer vision research, object detection has found practical applications in areas as diverse as autonomous vehicles, robots, security cameras, and people counting. Although deep neural networks have powerful feature representation capabilities in image processing and are often used as feature extraction modules in object detection, their introduction has changed conventional techniques of object identification and detection. Models trained using deep learning may be used as classifiers or regression tools, and they don't need any further human input. This suggests that object identification is a promising area for deep learning research. In order to recognize items in a picture, object detection first needs to pinpoint their exact position (object localization). [ABSTRACT FROM AUTHOR]

The development of aberrant cell proliferation in the lungs is a problematic condition that has the potential to result in death. On the list of diseases that most frequently result in mortality, lung cancer takes first place. The early stages of lung cancer are notoriously difficult to diagnose due to the fact that cancer cells with dimensions less than very small are notoriously difficult to spot by imaging. If the cell abnormalities are discovered in the early stages, it will be possible to begin therapy sooner, which will result in an improved chance of the patient surviving the illness. Several different image processing strategies can be utilized in the diagnostic phase of patient care to help spot signs of disease. In this paper, classification of Lung Cancer from chest X-ray images has been done using optimized Convolutional Neural Network (OCNN) and Ant Lion Optimization (ALO) algorithm. In pre-processing step, the contrast of all images are enhanced using Histogram Equalization (HE) method and the noises are removed from all images using Median Filtering. After the pre-processing step, feature extraction is performed using Gray Level Spatial Dependence (GLSD) to extract the statistical features. The feature vector is then trained and classified using OCNN-ALO algorithm. The ALO algorithm is used to optimize the hyper parameters of CNN layers. It classifies the lung images into normal and lung tumor affected. Performance results have indicated that OCNN-ALO attains the superior performance with 95.15% accuracy, 85.43% sensitivity, 93.4% specificity and 76.43% F1-score. [ABSTRACT FROM AUTHOR]

Semantic segmentation, once considered a daunting challenge in computer vision, has seen significant advancements with the rise of deep learning. This progress has transformed automated driving from a distant aspiration into a practical reality. However, many existing semantic segmentation algorithms were not specifically designed for autonomous vehicle applications; they were initially developed for more general image processing tasks. In this article, we introduce a reliable technique tailored for semantic segmentation in autonomous driving scenarios. It is imperative to swiftly develop a precise and up-to-date semantic segmentation system to ensure the safe and efficient operation of autonomous vehicles. At the core of this endeavor lies the concept of autonomous driving or self-driving cars. Without the capability to accurately identify and respond to hazards such as pedestrians, other vehicles, and traffic lanes, autonomous vehicles cannot function effectively. [ABSTRACT FROM AUTHOR]

This project is to build an Animal Intrusion Detection System that will alert the user or the respective authorities of sightings of animals in real time if it is detected entering the village or human establishment with the distance from the border. We will achieve this by training the model with the Mask RCNN algorithm. Mask R-CNN, sometimes known as Mask RCNN, is the most advanced Convolutional Neural Network (CNN) for instance and picture segmentation. Faster R-CNN, a region-based convolutional neural network, served as the foundation for Mask R-CNN. While training the model we will feed the animal photos with labelling. This will enable the detection and recognize the animal. This project was mainly developed while keeping elephants in mind. Since elephants are poached regularly for trading, meat, tusk, and entertainment purposes. Detecting animal intrusion using image processing helps the system send out alert messages to the user and respective authorities. This real-time implementation will help avoid and reduce animal-human accidents and save human properties from damage. This will also help the residents take quick action to find solutions. [ABSTRACT FROM AUTHOR]

Diabetic patients persist at risk of acquiring Proliferative Diabetic Retinopathy, a retinal condition (PDR). development of neovascularization, a disease in which aberrant blood vessels form on retina, one of key hallmarks of PDR. If not diagnosed & treated early enough, aforementioned illness preserve lead via blindness. Various image processing algorithms considering detecting neovascularization in fundus images have been proposed in a number of studies. Neovascularization, on other hand, difficult via identify due via its erratic growth pattern & tiny size. As a result of their ability via execute automatic feature extraction on objects among complex properties, deep learning approaches persist becoming more common in neovascularization recognition. A fore mentioned research proposes a transfer learning-based approach considering detecting neovascularization. We show certain suggested technique outperforms each individual in detecting neovascularization. [ABSTRACT FROM AUTHOR]

The use of Al enabled tools with smart data processing technologies has increased broadly in cites for the betterment and safety of the citizens. Data which we are talking here is of the data from video surveillance cameras on the street. It can record various human activities, using this feature we can record various anomalies which takes place like a Vehicle crash or human fall (due to sudden cardiac arrest, going faint, etc). Automatic accident detection system can secure the life of people who are in serious need of aid by providing timely treatment to the needy. This all thing can be done real time using a very famous object detection module known as YOLO. It uses fully convolutional neural network (CNN) along with OpenCV for the event prediction image processing. Under various circumstances and experimental data our system attained an accuracy of 80-86% in detecting vehicle crash and human fall, the average precision and fps was also increased using YOLOv4 as our underlying technology, the computational speed of the system also increased quite a bit when compared to CNN or R-CNN. Also, the immediate alert generation system which can send real-time alerts to the nearby authority becomes the most important aspect of this work. [ABSTRACT FROM AUTHOR]

Rice is a vital staple crop, and the presence of diseases can significantly impact its production. Consequently, there is an increasing demand to create efficient strategies for recognizing and controlling diseases in rice plants. This paper provides a comprehensive review of the recent advancements in disease identification techniques for rice plants, focusing on the integration of image processing, machine learning, and deep learning methodologies. Despite the progress made, there remains a need for further research to identify and classify additional rice plant diseases beyond the major ones discussed in this paper. By continually exploring new techniques and expanding the scope of disease identification, researchers have the potential to enhance the overall effectiveness of rice disease detection systems. This review serves as a valuable resource for researchers and practitioners in the field, offering insights into the current state of the art and suggesting future directions for advancements in the identification and management of rice plant diseases introduced. [ABSTRACT FROM AUTHOR]

The initial process of developing a Javanese script transliteration system to other scripts using a character recognition approach requires training data in the form of script images of all possible forms. The source of the dataset are script images from a book written in Javanese and then processed using image processing approach. The captured images were then grouped into their respective classes. The study starts with pre-processing the document images that includes the sub-processes of binarization, inverse, filtering, and followed by script segmentation using the projection profile method. Each script image is then processed in the feature extraction steps using the Intensity of Character or IoC algorithm. The feature data of each script image is then grouped using the K-Means clustering algorithm. The data was taken from the scan results of Hamong Tani's book on pages 2 and 59. After pre-processed and segmented images, 597 images of Javanese script were obtained. Using the IoC 3x3 feature, and the number of groups determined by 65 classes, the silhouette index value of the grouping results was found to be 0.5060. After calculating the ground truth value, it was found that the accuracy of the results was 86%. It can be concluded that the steps taken in this research can be used as a model in the process of providing a Javanese script images dataset. [ABSTRACT FROM AUTHOR]

In everyday life, new technologies in the area of image processing, mainly in domain of segmentation, emerge. Image segmentation is the process of dividing one image into several segments. It will smooth out the image and make it easier to analyze. The segmentation method also help you find the area of interest in an image. The purpose to become image simpler and more relevant. Existing image segmentation methods cannot handle every kind of images. The present review examines, evaluates, and reviews several picture segmentation approaches. [ABSTRACT FROM AUTHOR]

Underwater images often suffer from colour distortion, blurred details, low contrast, and a bluish or greenish tone due to the absorption and scattering of light in the underwater medium. These factors significantly impact visibility in underwater environments. Over the past few decades, numerous researchers have developed image restoration and enhancement methods to improve the quality and enhance the details of underwater images. This research paper provides an overview of various typical methods for underwater image restoration and enhancement which are thoroughly discussed. The evaluation aims to assess the performance and effectiveness of the methods. Finally, the research process of underwater image restoration and enhancement is summarized, and suggestions for future research directions are proposed to guide further advancements in this field. Underwater image processing is an emerging field of research with significant potential for aiding developers in exploring the underwater environment. This paper aims to provide insights into the degradation of underwater images, survey the state-of-the-art intelligence algorithms, for underwater image enhancement and restoration. It demonstrates the performance of various methods for underwater image dehazing and colour restoration and introduces a metric for evaluating the colour accuracy of underwater images. Furthermore, it offersan overview of significant applications of underwater imagery. [ABSTRACT FROM AUTHOR]

The main objective of image compression in the digital era is to reduce the size, of irrelevant and redundant data in digital images. this is able to keep and transfer the image in an efficient manner. In the present digital world, multimedia files have large sizes and consume a lot of space, and are also very difficult to transmit. Image compression plays a prominent role to reduce the size of the images and transmit such compressed data in a permitted time duration. This paper addresses image compression using machine learning methods. Machine learning is one of the subsets in artificial intelligence, in recent days machine learning plays prominent role in image processing. Using machine learning algorithms in image compression provides better-quality images and improves the different performance metrics. This paper attempts a study various machine learning algorithms as well as image compression techniques. [ABSTRACT FROM AUTHOR]

Various object detection models are presented in this paper. Definition of object is well defined in order to understand object detection models. Now-a-days the object detection which is computer vision and image processing related technology is widely used. Applications of object detections are infinite as it can be used all over like objects tracking, surveillance of videos, face recognition, detection of anomaly, counting of people, and the records still continue to many more. This paper compares various object detection algorithms the basic splitting is based on machine learning based and deep learning based. Fundamental objective of the paper focuseson Single stage object detectors (SSD) like You Only Look Once (YOLO) and Two stage object detectors as RCNN, Fast RCNN, etc. This paper's main goal is to highlight such object detection models and their use in various fields. [ABSTRACT FROM AUTHOR]

The data that was acquired by the globe Prosperity Affiliation reveals that the worldwide pandemic of the Covid virus has had a substantial impact on the globe and has now caused damage to more than eight million people all over the various regions of the world. Two of the most current health precautions are advised to be done immediately in order to avoid the spread of the contamination. The first is to wear facial coverings, and the second is to adopt safe social isolation practices. Both of these measures are recommended to be implemented immediately. We provide a productive strategy that is focussed on the continuous computerised observation of individuals in order to discriminate face coverings. This approach is based on the combination of TensorFlow and PC vision. The purpose of this action is to create a secure setting that makes a positive contribution to the overall health and happiness of the inhabitants of the community. Python is provided with an interpretation for the purpose of face mask detection as soon as the image processing system comes to the realisation that a breach has occurred. After then, an infrared sensor is kept in place in order to ascertain the total number of individuals present, supposing that the face covering is identified and that the door will be open. Under the premise that the number of people who are permitted to be inside the room is greater than the threshold, the door will not open. [ABSTRACT FROM AUTHOR]

The use of image processing to improve early diagnosis and treatment has been on the rise in several medical domains in the last few years. The high frequency of lung cancer and the difficulties in treating bronchogenic carcinoma have piqued the attention of the medical and scientific communities in the illness in recent years, particularly in situations when a rapid diagnosis is necessary for patients. Machine learning is used in cancer diagnostics to evaluate the patient's resistance, the probability of malignant growth recurrence and development, and the disease's permissiveness. In order to detect cancer in its early stages, the suggested framework uses computer tomography images of different stages of malignant development to predict its likely progression. The employment of a model was the most efficient and accurate technique to utilise the information to identify non-b malignant or malignant CT scan pictures. Furthermore, the model provided the most accurate and trustworthy predictions about the benign or malignant nature of CT scan pictures. [ABSTRACT FROM AUTHOR]

All traditional surgical procedures cause a lot of harm to the patient, while the biggest advantage of minimally invasive surgery is that the wound is small, the pain is small, and the recovery is fast. Recently, minimally invasive surgery has been widely adopted in the medical field and has been widely used as an important technique. Because of the high degree of coordination between the hand and the eye, there is a high demand for the operating ability of the physician. The images obtained by endoscopy do not have sufficient depth and are difficult to localize. The key to improving the visualization of such surgeries lies in optimizing image processing methods and feature point extraction algorithms that augment SLAM (Simultaneous Localization and Map Construction) techniques. This paper reviews several methods, such as the Contrast Constrained Adaptive Histogram Equalization (CLAHE) method, which solves the illumination problem by optimizing the grayscale transformation. Other methods such as wavelet transform, homomorphic filtering, and multiscale bilaterally weighted retina are also discussed for their image enhancement capabilities. For feature point extraction, AKAZE-ORB and SuperPoint algorithms combined with k-means algorithms are emphasized for their advancement in resolving uneven illumination and ensuring accurate extraction. [ABSTRACT FROM AUTHOR]

The scarecrow system described is a revolutionary crop protection approach combining tradition and innovation. It employs advanced features and user-friendly control to effectively deter animals and birds from damaging crops. The system utilizes sound-emitting speakers, disenchanting light beams, and hands with a flapping mechanism to startle and discourage intruders. The integration of time slots, a GSM module, and Raspberry Pi allows for convenient remote control and precise activation of the scarecrow. Additionally, a standalone image processing method with a camera module and Raspberry Pi detects and responds to animal presence, notifying the farmer and initiating deterrent actions. The system offers energy efficiency through time interval technology and utilizes solar-powered batteries for sustainability. Overall, the scarecrow system provides farmers with comprehensive crop protection and peace of mind. [ABSTRACT FROM AUTHOR]

Image processing is a developing area in communication of 2D signals, where in which image fusion gives a clear and important information or data about a particular image. This retrieval of all the information can be done using wavelet transform because the time and frequency domains are both employed in it. This paper deal with the mathematical formulation of the image fusion algorithm starting from image registration up to the resulting fused image. [ABSTRACT FROM AUTHOR]

Image pre-processing, also known as image enhancement, is a necessary first stage in basically all image processing applications. The quality of the sequent phases and the full application is highly dependent on outcome of the pre-processing stage. The removal of noise and soft tissue from CT images is an essential pre-processing step. During the filtering of both noise and soft tissue in images, the greater issue is maintaining location accuracy and image details without losing them. This paper describes an improved CT image pre-processing approach based on histogram modeling that reduces noise and undesirable artifacts and shows the external structure of the bone by taking image contents into account. Furthermore, the bone tissue is enhanced while the image texture's location accuracy is maintained. The powerful MATLAB software was utilized to test performance of the enhanced algorithm. In addition, the mean square error, peak signal to noise ratio, edge preservation index, and entropy are used to statistically compare the results and therefore evaluate their performance. When compared to the previous technique, the enhanced method produces better results. [ABSTRACT FROM AUTHOR]

Identification marks (or the fiducial markers) recognition is an important problem in image processing. This technique is used in robotics and industry to perform loading and unloading operations. Usually, algorithms using the neural networks to solve this problem show their low efficiency in devices with low computing resources. The proposed method is based on comparing the mark contour with an ellipse and converting the camera coordinates and contour parameters into the physical ones. So, it becomes possible to recognize the circular identification marks in space and to to perform the robot positioning without using large computing resources. The algorithm based on the proposed method was implemented on the Raspberry PI 3B quad core ARM Cortex-A53 in the 1.2 GHz single-board computer and an OV5647 camera. The algorithm showed acceptable positioning accuracy for solving tasks related to loading and unloading packaging units by autonomous robots without human intervention. [ABSTRACT FROM AUTHOR]

Glaucoma disease causes the loss of eye sight if unnoticed. Earlier detection and the extent to which the disease is progressed may provide good suggestions to the physicians for giving treatment. In this work the glaucoma severity stage with its disease types are understood the image processing techniques. In this work, an image segmentation algorithm, namely, OTSU's segmentation method is chosen. It detects optic cup (OC) and optic disc (OD). Several researchers found only Cup-to-Disk ratio (CDR). CDR did not apply to categorize the images whether they are normal or affected by glaucoma. So it is decided that calculating rim-to-disk ratio (RDR) may give an expected solution in assessing the glaucoma disease. The classifiers are used to find the progression of disease. In this research Support vector machine (SVM) is chosen as the classifier. The abnormal images are of four stages. They are Inferior, Superior, Nasal, Temporal and the types are found out by ISNT method. Further this work concludes the disease type by indicating either diabetic retinopathy or glaucoma. [ABSTRACT FROM AUTHOR]

A human-assisted manual classification might result in erroneous prediction and diagnosis, brain tumor identification is one of the most important and difficult challenges in the realm of medical image processing. Moreover, when there is a big number of data to be helped, it is difficult to work. Because brain tumors have such a wide range of appearances and because tumor and normal tissues are so comparable, extracting tumor areas from pictures is difficult. We suggested a method for extracting brain tumors from "magnetic resonance brain images (MRI)" using a Convolution neural network algorithm, followed by standard "convolutional neural network" categorization. The experiment used a real-time dataset. "Neural Network (CNN)" is a type of neural network that is created using Kera's and Tensor Flow and outperforms regular neural networks. Information on abnormal tissue development in the brain can be gleaned from MRIscanning. [ABSTRACT FROM AUTHOR]

Numerous approaches and practicals for detecting Alzheimer's disease have been developed over time. Understanding the scanned images of the human brain is the initial stage in this procedure. This explain the extent to which the brain has been damaged. The "Image Segmentation" software is used to examine the brain images. To extract usable information from an image, Image Processing employs a number of algorithms. Each algorithm generates output in its own way. As a result, the efficiency of each method differs. A few image segmentation algorithms are employed to segment the photos in order to obtain the needed information, and ground truth is used to validate the results. This validation is used to compare the segmented image's accuracy to the ground truth, indicating the algorithm's efficiency. The research focuses on U-net techniques, as well as the results obtained utilizing these algorithms. This result served as the foundation for the creation of an algorithm for detecting Alzheimer's disease. Because it is based on Alzheimer's disease, the entire procedure is based on 2-D brain pictures, and the outcome will be 2-D images that will be validated. Based on the portion of the brain to be segmented, brain MRI scans in various orientations are used as input. [ABSTRACT FROM AUTHOR]

In recent years, more information has been transferred in electronic form, so the science of cryptography got more attraction. Cryptography is an efficient and effective procedure to ensure that our data is accessed only by the intended receiver. Also, it makes sure that the data is not adulterated. In this paper, we have proposed hybrid cryptography for colour images which could provide a reliable resistance against differential security attacks. We have observed that the use of the blowfish algorithm and chaos logistic maps could ensure improved image security and performance, providing stable results in most conditions. [ABSTRACT FROM AUTHOR]

The purpose of this study is to improve the accuracy of biometric recognition systems by utilising innovative wavelet transform techniques as opposed to the more traditional Fourier Transform Technique. There are going to be two different teams working on this project. The Fourier Transform Technique is used for group 1, and the Novel Wavelet Transform Technique is used for group 2. Both groups have a sample size of 50. A matrix biometric database is combed through to obtain a sample of one hundred different scanned fingerprints. Using the MATLAB (R2017a) software tool, the values of accuracy are calculated to identify the performance of the novel wavelet transform technique. This is done with the help of image processing and also with image processing. Both the false acceptance ratio and the accuracy of the novel wavelet transform technique were found to be 2% and 82%, respectively, while those values for the fourier transform technique were found to be 6% and 76%, respectively. This value, 0.8, is derived from the G power. When compared to the fourier transform technique, it was found that the innovative wavelet transform method performed significantly better. The significance value p is greater than 0.05 but less than 0. In terms of accuracy, it would appear that the recently developed wavelet transform technique is a significant improvement upon its predecessor, the fourier transform technique. [ABSTRACT FROM AUTHOR]

Skin Cancer is one of the most common types of disease in the world, including Indonesia. This disease is caused by food that can be toxic to the human body and the effects of global warming. Especially melanoma, this type of skin cancer that occurs in melanocyte cells is a very serious threat. In the United States, although melanoma is only 5% of all cases of skin cancer that occur, melanoma has been the cause of death in 75% cases of death from skin cancer in the country. Melanoma has a similar shape as mole, birthmark or beauty sign that is often ignored. To recognize the presence of melanoma early on, this study uses the method of deep convolutional neural network to classify skin cancers that occur in melanocytes, where melanoma grows. The deep learning method used has good accuracy and good use in image recognition. Dermoscopy image of skin cancer is used as input for image processing. Before entering the classification stage, the image is processed with the image processing Scaling and contrast enhancement. After testing phase in this study, it concludes that the proposed method has the ability to classify skin cancer that occurs in melanocytes quite well with an accuracy of 93.3%. [ABSTRACT FROM AUTHOR]

Gastritis is a disease that occurs in the stomach and can be dangerous if not treated quickly. Gastritis can affect anyone, young or old. Therefore, a system is needed to identify early gastritis identification so that it can help the community to find out early on their gastritis. The identification method used in this study is the Radial Basis Function Neural Network with several image processing techniques. Tongue image is used as input image for image processing. Prior to the initial identification, the image preprocessing process is carried out, namely cropping, resizing, brightness, image feature extraction using color extraction using Hue Saturation Value (HSV). The dataset used in this study are limited to 150 tongue images and resulted in the ability to identify early gastritis with an accuracy of 93.3% for 50 testing images. [ABSTRACT FROM AUTHOR]