Search

The purpose of the research is to detect defects in textile industry using computer vision and fuzzy logic techniques. The paper presents various practices used for edge detection, including a fuzzy logic applied. By testing, a fuzzy logic method has been chosen to obtain a respectable result. The algorithm that is observed in the paper is improved by dividing the processing into two phases: a fast one, which allows us to assume that there is a defect, and a searching one, which allows a defect to be highlighted in color. The advanced algorithm was tested on several types of fabrics. An average background arrangement is obtained for them. That allows separating fabric texture and color from potential defects. The tests were also performed on fabrics (cloths) with different brightness, number and type of defects. We obtained the percentage ratios of the defect detection quality and so we suggested the ways to develop the system. [ABSTRACT FROM AUTHOR]

Technologies of Industry 4.0 are revolutionizing the way companies manufacture their products. Modern manufacturing systems are connected with the application of advanced sensors, increased automatization, embedded software, and robotics that collect and analyze data and allow for better decision-making. These trends are also reflected in recent innovations in the field of electrical discharge machining (EDM). One can state that recent progress in EDM is connected with the following areas: increase of automatization and robotization of the process, real-time data analysis, adaptive control and application of artificial intelligence, use of advanced sensors (especially machine vision systems), improvement of communication between machine tool and operator (i.e., voice-operated assistant technology) and application of additive manufacturing to shaping electrodes. All listed above and detailed innovations discussed in the paper cause EDM to become faster, more reliable, and more efficient technology, and the amount of unattended machining hours increases. [ABSTRACT FROM AUTHOR]

Image fusion is the integration of two or more images of the same scene taken at various times or by various sensors. To reduce the loss of contrast, texture, spectral, and spatial information in the fused image, we suggested an infrared (IR) and visible (VIS) image fusion technique employing a combination of (Latent Low Rank Representation) "LatLRR" and "ResNet." The fused image is more suitable and explicit for human and computer vision perception. Using LatLRR, the source images are initially broken down into saliency and latent low rank parts. Then, using weighted-average method, the Latent Low Rank components are integrated. We present a unique fusion outline based on deep features to overcome the problem of information lost by using RESNET to extract multilayer features for the saliency part as this network has a skip connection which helps to gather information from every layer. After that, the deep features are normalized by nuclear norm, and initial-weight maps are generated. Combining the original weight maps with a Soft-Max operation yields the final-weight maps. At last, by summing the fused Latent Low Rank component and the Saliency content, the fused image will be reconstructed. Both visual and quantitative performance evaluation is done for the proposed method with five existing methods and it is inferred from the results that the proposed method can well reconstruct the fused image without haloes and artifacts. [ABSTRACT FROM AUTHOR]

In the continuous development and burgeoning of AI powered technologies, the fruit classifier can benefit a lot of applications such as autonomous fruit harvesting and fruit disease detection assistant. Machine Learning is a revolutionary technique and has been a trending research and experimentation topic in the recent past years. It has been implemented to solve most day-to-day problems. However, there are more complex real-world challenges like computer vision and artificial intelligence that cannot be solved using normal machine learning techniques. In various fields, the Convolution Neural Networks (CNN) helps to classify the image and drive way to predict the accuracy rate of the classification. The detailed procedure on this work uses a deep learning technique for fruit classification. The fruit classification paves a very good example for a fine-grained image classification problem. Fine grained classification is challenging due to subtle differences among the images pertaining to each class. In the proposed CNN model, 6 primary classes, high quality fruit images with 81 classes, totally 57213 number of images are taken for training and evaluation procedure from Fruits 360 dataset. For unsupervised learning, GANs machine learning model has been used to predict more accuracy by creating new set of images that featured our own data. Dropout strategies are used to change the network itself in order to prevent overfitting and the SoftMax layer used as a activation function to ensure the input is to recorded for process or not. [ABSTRACT FROM AUTHOR]

Modern organizations face numerous physical security threats, from fire hazards to more intricate concerns regarding surveillance and unauthorized personnel. Conventional standalone fire and intrusion detection solutions must be installed and maintained independently, which leads to high capital and operational costs. Nonetheless, due to recent developments in smart sensors, computer vision techniques, and wireless communication technologies, these solutions can be integrated in a modular and low-cost manner. This work introduces Integrated Physical Security System (IP2S), a multi-agent system capable of coordinating diverse Internet of Things (IoT) sensors and actuators for an efficient mitigation of multiple physical security events. The proposed system was tested in a live case study that combined fire and intrusion detection in an industrial shop floor environment with four different sectors, two surveillance cameras, and a firefighting robot. The experimental results demonstrate that the integration of several events in a single automated system can be advantageous for the security of smart buildings, reducing false alarms and delays. [ABSTRACT FROM AUTHOR]

In recent years, improvements in natural language processing and computer vision have come together to provide automatic image caption generation. Image captioning is the process of creating a description for an image. Captioning an image needs the recognition of significant items, their properties, and their connections within the image. Additionally, it must create phrases that are syntactically and semantically accurate. Deep learning approaches can address the complexities and difficulties associated with image captions. This paper describes a joint model which is capable of automatically captioning images using EfficientNet-B0 and a transformer with multi-head attention. The model is an aggregation of an EfficientNet & Transformer single encoder and decoder. The encoder utilizes EfficientNet-B0, a convolutional neural network-based algorithm that generates a detailed input image, represented by embedding them into a fixed-length vector. The decoder employs a transformer, and a multi-head attention mechanism to selectively concentrate attention on certain regions of images to predict the sentence. The proposed model was trained using a large dataset Flickr8k to optimize BLEU N-Gram (N=1,2,3,4), METEOR score, and CIDEr, to assess the probability of the target description phrase given in the training images. Our studies show that the proposed model can produce captions for images automatically. [ABSTRACT FROM AUTHOR]

The agricultural industry is the most prominent sector for the economic growth of the country. Cotton is an important crop that brings out major returns to the economy. In the current era, smart agriculture technology is incorporated which includes various modern techniques such as sensors, artificial intelligence, and computer vision to digitalize the whole process. This digitalization through modern technology enhances the fast and intelligent productivity of crops. In this paper convolutional neuralnetwork-based methods have been used to predict the maturity of the cotton balls using image features of mature and immature images. Real-time maturity detection is a challenging task for cotton fiber because of the many constraints in natural light conditionsand environmental effects. The proposed scheme uses annotation to classify the cotton balls. The classification is labeled as matureand immature images then the annotated images are used for training the deep learning models Resnet 50 and VGG 16. The averageaccuracy achieved from these methods is 92.47% for Resnet 50 and 96.58% for VGG16. The loss of the model is 21.05% for Resnet50 and 14.91% for VGG16. [ABSTRACT FROM AUTHOR]

The paper discusses the prerequisites for the creation, technical solutions and implementation of an automated control system for the assembly of a small spacecraft. Both the hardware and software implementation of the system that provides control and logging of the assembly process of individual units at various workplaces are analyzed. The article presents solutions to reduce the requirements for equipment used to control the assembly technology, in particular, to use cameras with a lower resolution, through the use of special algorithms for the formation and processing of technological marks. A tool is presented that allows you to control the tightening torques of threaded connections and limit the tightening torque according to a given algorithm with wireless control. The developed system provides the functions of not only control, but also logging of the technological process, which can be useful in the future when creating a digital twin of the product. [ABSTRACT FROM AUTHOR]

Nowadays in India, due to the increase in the number of accidents vehicles are being automated using a variety of Computer vision algorithms. This paper focuses on the detection of tail signals of cars under different illumination circumstances. This system is implemented using FAST and SIFT algorithm which helps to extract features from the images. The obtained features were optimized by using K-Means Clustering Algorithm. This huge feature vector is converted into 8 clusters. These optimized feature vectors were trained on five different classifiers as Decision Tree, SVM(RBF), Random Forest, and Voting Classifier. The trained data set used in this algorithm contains around 9052 images. The obtained accuracy results of different classifiers are as follows, Decision tress has 76.36%, SVM(RBF) has 77.91%, Random Forest has 87.52%, KNN has 89.72%, the Voting classifier has 86.52%. It is observed that KNN gives the highest accuracy among the five used classifiers. This has, to the best of the authors' knowledge, not been presented in literature before. [ABSTRACT FROM AUTHOR]

Waste pollution in today's society is a significant contributor to environmental degradation. The detection of garbage bin overflow across the city is suggested to be accomplished by a combination of computer vision and machine learning. The system that is being proposed uses cropping and resizing in order to prepare images. In the suggested model, grey-level co-occurrence matrix (GLCM) and Haralick texture features are used for garbage detection. Five distinct GLCM feature vectors, Correlation, Homogeneity, Energy, Entropy, and Contrast, are chosen to characterize the texture of garbage. The Light gradient boosting machine (LGBM) is utilized for classification. The approach has a 93.71 percent accuracy rate. This strategy is more adaptable and makes less use of resources in comparison to the one that is currently being used for garbage identification. This technology not only enables accurate identification and recognition of garbage but also reduces operational costs by more than half, making it more cost-effective overall when compared to earlier monitoring approaches. [ABSTRACT FROM AUTHOR]

Face Emotion Detection Music Player is an innovative idea that combines computer vision and machine learning techniques to create a unique and interactive music player. This Research Paper studies the uses of Convolutional Neural Networks (CNN) to detect and analyze facial emotions in real-time and dynamically create playlists based on detected emotions. The music player is designed to record facial expressions using a webcam or other suitable camera. To reliably identify facial expressions including happiness, sorrow, rage, surprise, and more, a CNN model is trained on a sizable collection of face images. The trained model demonstrated an astounding 74.46% accuracy in identifying emotions when combined with music player software running on a PC or other suitable device. The CNN model analyses the photos to determine the user's emotional state as they interact with the music player in real-time, as captured by the camera. Based on the detected emotions, the player automatically selects and plays songs from a predefined music collection that matches the user's emotional state. For example, if the user looks happy, the player can play happy and energetic songs, while if the user looks angry, it can play more calming and soothing music. Face Detection Music Player offers a personalized and dynamic music listening experience as the music selection is constantly updated according to the user's emotions. This research paper illustrates how computer vision and machine learning techniques may be used to produce interactive programs that can instantly adjust to the moods and preferences of their users. It has uses in user experience research, entertainment, and mental health, among others. [ABSTRACT FROM AUTHOR]

Virtual Dressing Room (VDR) is an innovative technology that empowers users to attempt on attires virtually in a virtual atmosphere. The developed system eradicates the need for physical trial rooms at shopping store and delivers a continuous shopping experience. This system makes use of computer vision techniques to recognize and trace human body movements and overlaps a virtual clothing on the user's body. The proposed model is developed with the help of library imported from OpenCV and Python. The designed system encompasses of two major components, namely the human body detection and other one is the virtual garment overlay module. The user can interrelate with the developed model through a graphical user interface GUI) that permits the users to hand-picked outfits from a virtual wardrobe. The system's accuracy and performance were appraised on various test cases like specs, t-shirt, earring, pendants and the consequences showed promising results. The VDR system can transform the way people shop for clothes and provide a more immersive and convenient shopping experience. [ABSTRACT FROM AUTHOR]

The paper proposes a novel approach for recognizing and classifying gym exercises using human pose detection with Mediapipe and LSTM, followed by matching the accuracy of the exercises using the Hungarian algorithm and cosine similarity. The proposed approach involves the use of a combination of cutting-edge technologies, including computer vision and deep learning, to recognize and classify a range of gym exercises accurately and efficiently. This study presents several advantages of this approach, including the ability to leverage temporal information, robustness to variations in lighting, camera angle, and occlusions, and the ability to handle a wide variety of gym exercises. Experimental results demonstrate that the proposed approach achieves high accuracy and outperforms existing methods for gym exercise recognition and classification. The paper also compares the proposed approach with existing techniques, highlighting its strengths and cost-effectiveness. It concludes by outlining the methodology, including dataset collection and pre-processing, landmark extraction, LSTM model training, exercise classification, and similarity analysis. Finally, the paper discusses potential future research areas in this field. [ABSTRACT FROM AUTHOR]

This state-of-the-art review paper provides a comprehensive overview of meta-learning techniques for learning to learn in machine learning. Meta-learning called "learning to learn," is an approach that enables models to learn from new tasks learning quickly and efficiently, based on prior knowledge and experience. The paper discusses various meta-learning techniques, such as model-based meta-learning, metric-based meta-learning, optimization-based meta-learning, and memory-based meta-learning. The review also covers the applications of meta-learning in different domains, including natural language processing, computer vision, and reinforcement learning. Additionally, the paper discusses the benchmark datasets, challenges along with the future directions of meta-learning research, highlighting the need for developing more efficient and scalable meta-learning algorithms. Moreover, the proposed Meta-Learn framework incorporates a hierarchical structure to facilitate knowledge transfer across multiple levels of abstraction, and it also includes a memory component that enables the system to store and reuse previous experiences. Overall, this review aims to provide a comprehensive overview of Meta-Learning techniques for learning to learn in Machine Learning and proposes a novel Meta-Learn framework that could potentially enhance the performance and scalability of Meta-Learning algorithms. [ABSTRACT FROM AUTHOR]

Human-computer interaction relies heavily on the ability to understand and respond to human signs, and the interpretation of hand gestures is a fundamental part of this. It's because there's a strong need to make interactions between people and machines feel more natural. This is due to the intense desire to make communication between humans and computers or other devices natural. This recognition makes it possible for computers to capture and understand hand motions. Recognizing hand gestures under computer vision is crucial for facilitating communication between the deaf and dumb community and normal people, as well as between the elderly who cannot wear hand gloves or sensors and their caregivers. As a result, there is an urgent need for a system with accuracy and the ability to recognize these gestures, as the problem of inaccurately recognizing hand gestures may hurt the human community that relies on gestures to transmit their desires. Given the importance of applications for hand gesture recognition, a completely new method has been developed for recognizing hand gestures and applying it to the hand gesture data set. The study's goal is to illuminate the most crucial steps in the hand gesture detection process. which is the process of detection and recognition of hand gestures by using hand landmark points for tracking hands and using the multi-connected architecture of associative memory as a new trend in the recognition phase. The proposed method gave high accuracy in real-time and showed promising outcomes, with accuracy for American, Chinese, and Arabic sign language and numbers at 95.42%, 92.13%, 93.55%, and 94%, respectively. [ABSTRACT FROM AUTHOR]

Oleoresin is a non-timber forest product of Dipterocarpus spp. tree from Dipterocarps family that can be yield from tapping process. Some countries, such as Thailand, and Laos, the combustion method was conducted in the tapping hole to increase oleoresin yield. However, in East Kalimantan, the combustion method in tapping holes has never been carried out, and results are still unknown. This study aimed to determine the effect of the combustion method in Dipterocarpus grandiflorus tapping hole on volume, specific gravity and color changes of oleoresin yielded. The research was carried out in aforest area with particular purposes Labanan, Berau, East Kalimantan by making burning and non-burning tapping holes in two groups of different diameters of D. grandiflorus trees. The volume of oleoresin yielded for two weeks was collected in a plastic bottle and then measured. Specific gravity testing was conducted using the pycnometer method and the color properties of oleoresin by computer vision analysis method. The present study showed that a tree diameter of 50 cm contained more oleoresin yield than a tree diameter of 40 cm, both in combusted and non-combusted tapping holes. The combustion method showed a decreased on volume of oleoresin up to 39.03%, but the specific gravity increased to 0.96-0.97. Meanwhile, the combusted method produced more darker oleoresin, namely L* 11.75, a* 3 and b* 7.75. However, this recent study provided that the best method to obtained oleoresin is non-combusted tapping holes. [ABSTRACT FROM AUTHOR]

The detection of image edges is a major challenge in image processing and computer vision, since it needs to dispense of unnecessary pixels while preserving significant ones which indicate the image's structure. An optimized algorithm for image edge detection is introduced in this study, which uses a Genetic Optimization (GO) Algorithm to seek out sources of the maximum pixel value to optimally allocate the image edges. The algorithm simulates the evaluation of various values and consists of three operations: selection, crossover operations, and thresholding the optimized value. The community is made up of these evaluations which are conducted at each stage to select the fittest individuals and eliminate the less fit ones. By repeating this process, GO eventually identifies the image edges and eliminates non-edge pixels. The results of the study show that the proposed algorithm successfully addresses the challenges associated with image edge detection. The limitations of the research are also discussed, and potential areas for future research are suggested. This algorithm is expected to be useful for solving complex problems. [ABSTRACT FROM AUTHOR]