Search

AbstractBased on the adverse physiological and social consequences associated with sadness, recognizing this emotion is crucial to mitigate its negative impacts. To address this issue, an efficient system was proposed for recognizing different level of sadness using embedded technology. The system captured speech input through a microphone and then processed it on Raspberry Pi microcomputer. The key feature extraction technique used was Bark-Frequency Cepstral Coefficient (BFCC) method. After selecting a set of features, a thorough analysis of the computational algorithms was conducted, and dimensionality reduction method was applied to reduce computational costs, including noise. These features served as a pre-processing stage for Machine Learning (ML) model based on K-Nearest Neighbors (KNN) architecture. The results were then shown on a smartphone application through Bluetooth connectivity. Based on experimental results, the system showed an impressive improvement of 11.97 dB in the signal-to-noise ratio (SNR) compared to feature extraction using Mel-frequency Cepstral Coefficient (MFCC) method. The results also showed a favorable accuracy rate, implying the effectiveness of KNN with a k-value of 5 in accurately identifying different level of sadness. The system particularly excelled in correctly recognizing high and low level of sadness. Therefore, the effective categorization of sadness emotion level could be applied to the early diagnosis of mental disorders such as depression and anxiety, contributing to the reduction of the adverse impact of sadness. [ABSTRACT FROM AUTHOR]

Copyright of Engineer Military Bulletin / Muszaki Katonai Közlöny is the property of National University of Public Service and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

This study introduces an advanced biometric recognition system that seamlessly integrates facial and palm print modalities, showcasing outstanding performance across diverse datasets. The facial recognition dataset exhibits remarkable training, validation, and test accuracies at 98.98%, 98.99%, and 99%, respectively. Precision, recall, and F1-score metrics consistently reach 99%, underlining the system's robust and reliable performance in facial identification. Similarly, the palm print dataset demonstrates impressive results, with training, validation, and test accuracies at 98.90%, 99%, and 99%, respectively. Precision, recall, and F1- scores maintain a high level of 99%, emphasizing the system's effectiveness in palm print recognition. The combined "Face and Palm" dataset further highlights the system's exceptional capabilities, achieving perfect scores of 100% in training, validation, and test accuracies, as well as precision, recall, and F1-score. This underscores the system's versatility and proficiency in simultaneously recognizing facial and palm features. The innovative fusion of facial and palm print modalities in this biometric recognition system yields impressive and consistent results across multiple datasets. The system's high accuracy and precision, coupled with its adaptability to various scenarios, position it as a valuable advancement in biometric technology. [ABSTRACT FROM AUTHOR]

Copyright of Engineer Military Bulletin / Muszaki Katonai Közlöny is the property of National University of Public Service and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Big data and intelligence artificial have greatly broadened the channels for people to acquire knowledge. The practical application of artificial intelligence technology in college English classrooms can not only help teachers prepare more scientific and advanced teaching plans, and also effectively improve the quality of college English teaching. In this paper, we employ deep neural networks to design a big data-driven English teaching improvement scheme. The proposed method can help teachers and students improve teaching strategies and assist students to learn more accurately and efficiently. First, we analyze the current deficiencies and reasons in English teaching. Second, we analyze the application of artificial intelligence in text corpus construction. Third, we use neural networks to design an intelligent face recognition system for teachers to improve the attendance rate. Forth, by analyzing the students' writing, we design a writing scoring method based on the neural network model, which can evaluate students' writing content. Finally, we experimentally evaluate and verify the effectiveness of the proposed method in this paper. [ABSTRACT FROM AUTHOR]

Background: Due to the uniqueness of individual fingerprint patterns, fingerprint recognition is a critical biometric identifying tool. Because of its pervasiveness and dependability, this technology requires ongoing refining and comprehension of its core concepts and applications in diverse real-world circumstances. Objective: This article aims to show how to use Gabor filters and the Euclidean distance to extract and match features. The system aims to explain the fundamental ideas of fingerprint recognition and show how accurate and efficient it is at identifying people. Methods: The system uses Gabor filters to capture fine details in fingerprint images, producing a fixed-length finger code. The Euclidean distance is employed to compare finger codes, making the matching process more efficient. The suggested algorithm is rigorously tested using the FVC2000 database, with picture quality improvements such as Gaussian filtering and histogram equalization implemented. Results: The system demonstrated its proficiency in fingerprint identification with a significant accuracy of 98.22%. The results highlight the system's basic character while recognizing the need for sophisticated methodologies and application adjustments. Conclusion: While the system serves as a core model for comprehending fingerprint identification, it emphasizes the significance of future study and development in picture pretreatment, feature extraction methods, matching approaches, and database administration. This article's insights contribute to the continuous advances in biometric identification technology by resolving problems and improving the reliability of fingerprint recognition systems.

India has more than 1,600 official languages, making it a multilingual country. Kannada, one of the major languages, originated in the state of Karnataka and is currently ranked 33rd among the accents that are most often spoken throughout the world. However, the survey shows that much more effort is needed to create a complete handwritten identification system. Segmentation is one of the crucial steps in a handwriting identification system that extracts significant objects from an image. The feature extraction and classification phases of handwritten text recognition will be more successful if the segmentation approaches selected are efficient. In the proposed system, segmentation was accomplished using bounding box and contour tracing methods. The result got is delivered to the next step of handwritten identification system. An average accuracy of 92.6% is worked out for line segmentation and word segmentation. [ABSTRACT FROM AUTHOR]

Video surveillance system is widely used in modern society. Human behavior recognition based on neural network uses computer vision and image processing technology to analyze and judge the behavior of different objects in surveillance video. In order to improve the recognition accuracy of human behavior, this paper makes an in-depth study on the design of video surveillance system and target tracking using the relevant algorithms of neural network. In this paper, the application of relevant neural network algorithms and the technology of character behavior recognition are analyzed by means of experimental method comparison and algorithm comparison. The experimental data shows that in the detection of abnormal behavior, the highest accuracy rate of human behavior recognition is 90% by using the improved hu invariant moment model. [ABSTRACT FROM AUTHOR]

Copyright of Engineer Military Bulletin / Muszaki Katonai Közlöny is the property of National University of Public Service and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

【Objective】A practical mobile APP recognition system based on Android was designed for the requirement of real-time detection of various apple bark diseases in orchards.【Method】The images of ring rot, canker and dry rot were collected through network searching and field shooting. After amplification and labeling, the training set and test set were divided according to the ratio of 8 ∶ 2. The YOLOv5s algorithm was used to train the apple bark disease recognition network model. The lightweight network model trained was deployed on the Android end, and the corresponding APP interface was designed to realize the rapid diagnosis of ring rot, canker and dry rot.【Result】The recognition effect of the deep learning network obtained after training is good, the accuracy rate is stable at 88.7%, the recall rate is stable at 85.8%, and the average accuracy value is stable at 87.2%. Among them, the accuracy of canker is 93.5%, dry rot is 88.2%, and ring rot is 84.3%. After it is deployed on the Android end, the processing time of each disease picture is less than 1s, and the detection confidence is 87.954%. The lightweight recognition system not only realizes the rapid detection of the three diseases, but also ensures high recognition accuracy.【Conclusion】The YOLOv5s network weight model is small, which can be easily deployed on the Android. The APP designed based on YOLOv5s is simple to operate with high detection accuracy and fast recognition speed, which is helpful for precise management of orchards.

People with autistic spectrum disorders (ASDs) have difficulty recognizing and engaging with others. The symptoms of ASD may occur in a wide range of situations. There are numerous different types of functions for people with an ASD. Although it may be possible to reduce the symptoms of ASD and enhance the quality of life with appropriate treatment and support, there is no cure. Developing expert systems for identifying ASD based on the facial landmarks of children is the main contribution for improvements in the healthcare system in Saudi Arabia for detecting ASD at an early stage. However, deep learning algorithms have provided outstanding performances in a variety of pattern-recognition studies. The use of techniques based on convolutional neural networks (CNNs) has been proposed by several scholars to use in investigations of ASD. At present, there is no diagnostic test available for ASD, making this diagnosis challenging. Clinicians focus on a patient's behavior and developmental history. Therefore, using the facial landmarks of children has become very important for detecting ASDs as the face is thought to be a reflection of the brain; it has the potential to be used as a diagnostic biomarker, in addition to being an easy-to-use and practical tool for the early detection of ASDs. This study uses a variety of transfer learning approaches observed in deep CNNs to recognize autistic children based on facial landmark detection. An empirical study is conducted to discover the ideal settings for the optimizer and hyperparameters in the CNN model so that its prediction accuracy can be improved. A transfer learning approach, such as MobileNetV2 and hybrid VGG19, is used with different machine learning programs, such as logistic regression, a linear support vector machine (linear SVC), random forest, decision tree, gradient boosting, MLPClassifier, and K-nearest neighbors. The deep learning models are examined using a standard research dataset from Kaggle, which contains 2940 images of autistic and non-autistic children. The MobileNetV2 model achieved an accuracy of 92% on the test set. The results of the proposed research indicate that MobileNetV2 transfer learning strategies are better than those developed in existing systems. The updated version of our model has the potential to assist physicians in verifying the accuracy of their first screening for ASDs in child patients. [ABSTRACT FROM AUTHOR]

Copyright of Engineer Military Bulletin / Muszaki Katonai Közlöny is the property of National University of Public Service and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

This paper presents the development of a system to automatically recognize Nigerian traditional attire worn by major ethnic groups. Samples of clothing images depicting the cultural groups were obtained from publicly available online sources and an architecture to classify the images classes was designed using the Convolutional Neural Network (CNN) model. Performance evaluation results from experiment show that the model can classify images accordingly, achieving a validation accuracy score of 86 %. A web application interface was also implemented to validate the model's accuracy which shows good potential when integrated in commercial clothing applications. [ABSTRACT FROM AUTHOR]

Injuries and the associated surgery to the anterior cruciate ligament (ACL) can often trigger unpredictable effects, such as the so-called giving way effect, which is an uncontrolled buckling of the knee joint. For this purpose, the KneTex project has developed a smart textile-integrated sensor and actuator bandage system to record the movement of such patients and to monitor and support the rehabilitation process. Long-term monitoring and analysis of the movement data will identify patterns or gait types that can lead to a giving way effect. This paper describes the recent developments of the random forest model-based motion classification system developed within the project. Improvements have been achieved by reducing the number of features needed by 25% using feature importance analysis, speeding up the computation time by 14%, and increasing the classification efficiency. Feature elimination is a useful tool to improve classification systems in settings where feature count is high and feature importance analysis contributed by improving our understanding which sensor of our system are important for the motion classification task.

As an essential basic function of grassland resource surveys, grassland-type recognition is of great importance in both theoretical research and practical applications. For a long time, grassland-type recognition has mainly relied on two methods: manual recognition and remote sensing recognition. Among them, manual recognition is time-consuming and laborious, and easily affected by the level of expertise of the investigator, whereas remote sensing recognition is limited by the spatial resolution of satellite images, and is not suitable for use in field surveys. In recent years, deep learning techniques have been widely used in the image recognition field, but the application of deep learning in the field of grassland-type recognition needs to be further explored. Based on a large number of field and web-crawled grassland images, grassland-type recognition models are constructed using the PyTorch deep learning framework. During model construction, a large amount of knowledge learned by the VGG-19 model on the ImageNet dataset is transferred to the task of grassland-type recognition by the transfer learning method. By comparing the performances of models with different initial learning rates and whether or not data augmentation is used, an optimal grassland-type recognition model is established. Based on the optimal model, grassland resource-type map, and meteorological data, PyQt5 is used to design and develop a grassland-type recognition system that uses user-uploaded grassland images and the images' location information to comprehensively recognize grassland types. The results of this study showed that: (1) When the initial learning rate was set to 0.01, the model recognition accuracy was better than that of the models using initial learning rates of 0.1, 0.05, 0.005, and 0.001. Setting a reasonable initial learning rate helps the model quickly reach optimal performance and can effectively avoid variations in the model. (2) Data augmentation increases the diversity of data, reducing the overfitting of the model; recognition accuracies of the models constructed using the augmented data can be improved by 3.07–4.88%. (3) When the initial learning rate was 0.01, modeling with augmented data and with a training epoch = 30, the model performance reached its peak—the TOP1 accuracy of the model was 78.32% and the TOP5 accuracy of the model was 91.27%. (4) Among the 18 grassland types, the recognition accuracy of each grassland type reached over 70.00%, and the probability of misclassification among most of the grassland types was less than 5.00%. (5) The grassland-type recognition system incorporates two reference grassland types to further improve the accuracy of grassland-type recognition; the accuracy of the two reference grassland types was 72.82% and 75.01%, respectively. The recognition system has the advantages of convenient information acquisition, good visualization, easy operation, and high stability, which provides a new approach for the intelligent recognition of grassland types using grassland images taken in a field survey. [ABSTRACT FROM AUTHOR]

Identity documents recognition is far beyond classical optical character recognition problems. Automated ID document recognition systems are tasked not only with the extraction of editable and transferable data but with performing identity validation and preventing fraud, with an increasingly high cost of error. A significant amount of research is directed to the creation of ID analysis systems with a specific focus for a subset of document types, or a particular mode of image acquisition, however, one of the challenges of the modern world is an increasing demand for identity document recognition from a wide variety of image sources, such as scans, photos, or video frames, as well as in a variety of virtually uncontrolled capturing conditions. In this paper, we describe the scope and context of identity document analysis and recognition problem and its challenges; analyze the existing works on implementing ID document recognition systems; and set a task to construct a unified framework for identity document recognition, which would be applicable for different types of image sources and capturing conditions, as well as scalable enough to support large number of identity document types. The aim of the presented framework is to serve as a basis for developing new methods and algorithms for ID document recognition, as well as for far more heavy challenges of identity document forensics, fully automated personal authentication and fraud prevention.

The unrestricted consumption of fossil fuels negatively impacts the economic, social and environmental aspects, observed from a sustainable perspective. Therefore, it is necessary to develop and adopt skills that enable the monitoring and mitigation of risks to the environment. In view of this, we propose a method with multiple approaches emphasizing a three-dimensional perspective of energy consumption by diesel engines, which represent one of the main pollutants emitters in transport. As a methodology, fuzzy logic was adopted, together with a recognition system, in order to mitigate the uncertainties inherent to the applied data. The procedure was applied to the city of Rio de Janeiro, Brazil, with information collected from seven toll plazas and a radar with volumetric counting. The results indicate a good adherence of the sustainability index to real cases, allowing a better observation of changes in environmental criteria and a more efficient inspection in the application of good practices, in addition to enabling greater participation of society in the inspection and adoption of environmental criteria in transport. [ABSTRACT FROM AUTHOR]

Aim: The aim is to improve and develop a written digit identification to detect practically important issues in pattern recognition using Novel Convolution Neural Networks compared with Decision tree. Materials and Methods: Handwritten digit recognition is performed using Novel Convolution neural networks algorithm Sample size is 66 over decision tree algorithm Sample size is (N=66) with the split size using Table 1 and training using G power 80% and testing dataset 70% and 30% respectively. Results: The retrieval accuracy of the Novel Convolution Neural Networks classifier is (96.42%) and Decision tree is (72.35%), There exists a statistically insignificant difference between the two groups (p=0.193; p>0.05)`. Conclusion: The work has confirmed that the efficiency of the Novel convolution neural network algorithm has given more accuracy value in written digit identification when compared to decision tree algorithm using artificial intelligence algorithm. [ABSTRACT FROM AUTHOR]

Nowadays, high security is the most essential requirement for spoofing attacks. There is more demand for more security and more precision and more speed of authentication for buying electronics. There is a large scope solution for security issues is human observable and physiology characteristic in bios crypt. In any case, in phrases of both time and space the biometric systems are highly complex, for high security this is not that much appropriate. Consequently, for high authentication we proposed a finger vein recognition system that one give high verification contrasted with other existing strategies. [ABSTRACT FROM AUTHOR]

After surgery of the anterior cruciate ligament (ACL), unpredictable effects can occur in the patient’s gait patterns. Thus, patients often experience instability and uncontrolled buckling of the knee joint (“giving way” effect). The “giving way” effect is also difficult to measure due to the uncontrolled and rare occurrence. Within the KneTex project a smart textile-integrated sensor and actuator bandage system is developed to detect instability of the knee joint and actively support the rehabilitation process. Based on the sensor data the calculation of the knee angles as well as a recognition system is implemented. The bandage system can be controlled via a custom app which is developed as part of the project. This paper describes the recent developments and current status of the project. In addition, the results and steps for future work are described as well.

Sign language is used by people who have hearing and speaking difficulties, but not understood by many without these difficulties. Therefore, sign language recognition systems are developed to aid communication between hearing impaired people and others. This paper developed a static American Sign Language Recognition (ASLR) system using canny-edge and histogram of oriented gradient (HOG) for feature extraction with K-Nearest Neighbour (K-NN) as classifier. The sign language image datasets used consist of English alphabets from both Massey University and Kaggle, and numbers (0-9) from Massey University. Median filter was used to remove noise after images were converted to grayscale. Otsu algorithm was used for segmentation while edges in the images were preserved using canny edge detection technique with HOG parameters tuning to obtain feature vectors. The extracted features were used by K-NN for classification. An average recognition accuracy and computational testing time of 97.57% and 0.39s respectively were obtained based on experiments with the Massey University dataset. Similarly, an average recognition accuracy and computational testing time of 98.97% and 0.43s respectively were obtained based on experiments with the Kaggle dataset. The developed system successfully recognized static English alphabets and numbers and outperformed some existing systems.

This paper proposes a deep learning approach to recognize Tamil Letter from images which contains text. This is recognition process, the text in the images are divided to letter or characters. Each recognized letters are sending to recognition system and filter the text using deep learning algorithms. Our proposed algorithm is used to separate letter from the text using convolution neural network approach. The filtering system is used for identifying font based on that letters are found. The Tamil letters are test data and loaded in recognition systems. The trained data are input which contains filtered letter from image. For example, Tamil letters such as are available in test dataset. The trained data are applied into deep convolution neural network process. The two dataset are created which contains test data with Tamil letter and second one for recognized input data or trained data. 15 thousands of letters are taken and 512 X 512 X 3 size deep convolution network is created with font and letters. As the result, 85% Tamil letters are recognized and 82% are tested using font. TensorFlow is used for testing the accuracy and success rate.

People with autistic spectrum disorders (ASDs) have difficulty recognizing and engaging with others. The symptoms of ASD may occur in a wide range of situations. There are numerous different types of functions for people with an ASD. Although it may be possible to reduce the symptoms of ASD and enhance the quality of life with appropriate treatment and support, there is no cure. Developing expert systems for identifying ASD based on the facial landmarks of children is the main contribution for improvements in the healthcare system in Saudi Arabia for detecting ASD at an early stage. However, deep learning algorithms have provided outstanding performances in a variety of pattern-recognition studies. The use of techniques based on convolutional neural networks (CNNs) has been proposed by several scholars to use in investigations of ASD. At present, there is no diagnostic test available for ASD, making this diagnosis challenging. Clinicians focus on a patient’s behavior and developmental history. Therefore, using the facial landmarks of children has become very important for detecting ASDs as the face is thought to be a reflection of the brain; it has the potential to be used as a diagnostic biomarker, in addition to being an easy-to-use and practical tool for the early detection of ASDs. This study uses a variety of transfer learning approaches observed in deep CNNs to recognize autistic children based on facial landmark detection. An empirical study is conducted to discover the ideal settings for the optimizer and hyperparameters in the CNN model so that its prediction accuracy can be improved. A transfer learning approach, such as MobileNetV2 and hybrid VGG19, is used with different machine learning programs, such as logistic regression, a linear support vector machine (linear SVC), random forest, decision tree, gradient boosting, MLPClassifier, and K-nearest neighbors. The deep learning models are examined using a standard research dataset from Kaggle, which contains 2940 images of autistic and non-autistic children. The MobileNetV2 model achieved an accuracy of 92% on the test set. The results of the proposed research indicate that MobileNetV2 transfer learning strategies are better than those developed in existing systems. The updated version of our model has the potential to assist physicians in verifying the accuracy of their first screening for ASDs in child patients.

In this age of artificial intelligence, facial expression recognition is an essential pool to describe emotion and psychology. In recent studies, many researchers have not achieved satisfactory results. This paper proposed an expression recognition system based on ResNet-152. Statistical analysis showed our method achieved 96.44% accuracy. Comparative experiments show that the model is better than mainstream models. In addition, we briefly described the application of facial expression recognition technology in the IoT (Internet of things). [ABSTRACT FROM AUTHOR]

Automatic license plate recognition has become a significant tool as a result of the development of smart cities. During the experiment studied in the current paper, 50 images were used to detect Saudi car plates. After the preprocessing stage, the canny edge method to detect the car edges and different threshold techniques were used to reduce noise. Horizontal projection was applied in the segmentation process to split the plate. After that, a masking technique was utilized to locate and separate the region of interest in the image. OCR was applied to the processed images to read the characters and numbers in English and Arabic separately. Then, combining the English and Arabic text, after using the re-shaper for the Arabic letters. Finally, rendering of the results of text on images down the plate regions took place. The canny algorithm with projection technique with a proper preprocessing for images produces results with accuracy of 92.4% and 96% for Arabic and English language respectively. [ABSTRACT FROM AUTHOR]

Stair region detection and recognition from a stair candidate image is a challenging work in the computer vision research area. In the last few decades, researchers use many recognition systems to recognize and verify the stair region from other analogous objects. However, all the verification systems such as vanishing point (VP) do not achieve the desired result for various reasons. In this regard, a method is proposed in this paper to investigate the vanishing point’s problem arising in the case of stair region verification based on the three basic criteria, i.e. focal angle of the camera, height of the camera from the ground, and distance of the camera from the stair image. For that, primarily, the stair region is extracted by utilizing the geometrical features of a stair. The detected stair candidate region is verified through the y coordinate value of the vertical VP, i.e.y

Hand gesture recognition system helps in development of interface system for entering text in human computer interaction. In this paper, we have presented a hand gesture recognition system designed for dataset consisting of numerals and alphabets in lower case. The proposed system detects the hand with the help of skin color and motion information. Hand tracking is done with the help of two-level tracking system using modified Kanade–Lucas–Tomasi (KLT) tracking algorithm. The existing KLT was not able to track the gesture trajectory once the skin detected becomes less in area resulting in decreased number of points. In this paper, traditional KLT has been modified with a new additional feature to overcome this difficulty. In feature extraction process, a feature matrix consisting of 30 features have been created. Among these 30 features, few features like density-1, density-2, and perimeter efficiency have been introduced and are used for calculating efficiency along with some existing features. Inclusion of new features helps in improving the performance and accuracy of the system. Recognition is done using six classifiers including SVM (Support Vector machine), Decision Tree, Naïve Bayes, k-NN (K nearest neighbor), ANN (Artificial neural Network) and ELM (Extreme learning Machine). The experimental results prove that 89.67% of accuracy is achieved for the recognition of dataset containing both numerals and alphabets. Our proposed system is also compared with two existing literatures and it has been observed that better accuracy is exhibited by the proposed system. [ABSTRACT FROM AUTHOR]