Automatic Recognition Technology Research Articles

Deep learning-based anatomical position recognition for gastroscopic examination.

The gastroscopic examination is a preferred method for the detection of upper gastrointestinal lesions. However, gastroscopic examination has high requirements for doctors, especially for the strict position and quantity of the archived images. These requirements are challenging for the education and training of junior doctors. The purpose of this study is to use deep learning to develop automatic position recognition technology for gastroscopic examination. A total of 17182 gastroscopic images in eight anatomical position categories are collected. Convolutional neural network model MogaNet is used to identify all the anatomical positions of the stomach for gastroscopic examination The performance of four models is evaluated by sensitivity, precision, and F1 score. The average sensitivity of the method proposed is 0.963, which is 0.074, 0.066 and 0.065 higher than ResNet, GoogleNet and SqueezeNet, respectively. The average precision of the method proposed is 0.964, which is 0.072, 0.067 and 0.068 higher than ResNet, GoogleNet, and SqueezeNet, respectively. And the average F1-Score of the method proposed is 0.964, which is 0.074, 0.067 and 0.067 higher than ResNet, GoogleNet, and SqueezeNet, respectively. The results of the t-test show that the method proposed is significantly different from other methods (p< 0.05). The method proposed exhibits the best performance for anatomical positions recognition. And the method proposed can help junior doctors meet the requirements of completeness of gastroscopic examination and the number and position of archived images quickly.

Utilizing Multiple Regression Analysis and Entropy Method for Automated Aesthetic Evaluation of Interface Layouts

Aesthetic evaluation of increasingly complex and personalized human–computer interaction interfaces serves as a critical bridge between humans and machines, fundamentally enhancing various interaction factors. This study addresses the challenges in aesthetic evaluation by adjusting existing methodologies to incorporate seven aesthetic metrics: density, symmetry, balance, proportionality, uniformity, simplicity, and sequence. These metrics were effectively integrated into a composite evaluation metric through both multiple regression analysis and entropy methods, with the efficacy of both fitting methods validated. Leveraging automatic segmentation and recognition technology for interface screenshots, this research enables rapid, automated acquisition of evaluations for the seven metrics and the composite index, leading to the development of a prototype system for interface layout aesthetic assessment. Aimed at reducing the time, manpower, and resources required for interface evaluation, this study enhances the universality, compatibility, and flexibility of layout assessments. It promotes integration at any stage of the design process, significantly benefiting lightweight rapid evaluation and iterative design cycles, thereby advancing the field of interface aesthetic evaluation.

StreetSavvy: Automated Parking Management System with License Plate Recognition and Pricing

The Parking Charge Calculator represents a meticulously designed system aimed at accurately determining the parking charges for vehicles within a parking facility. This innovative solution leverages Automatic Number Plate Recognition (ANPR) technology in conjunction with a robust database system to streamline the process of identifying vehicles and computing corresponding parking fees. Within the confines of a parking lot, the unique identification of vehicles is facilitated through the extraction of information from license plates. When a vehicle enters the parking lot, the Parking Charge Calculator initiates the retrieval of the license plate number through the ANPR technology. This automated recognition system efficiently captures and records the alphanumeric sequence on the license plate, laying the foundation for subsequent charge calculations. The fundamental mechanism employed by the Parking Charge Calculator involves the multiplication of hourly parking charges by the temporal duration of a vehicle's stay. As the vehicle enters the parking lot, the system records the entry time. Upon the vehicle's departure, the departure time is similarly recorded. The time difference between these two instances becomes a critical parameter in the calculation process. Hourly parking charges, predetermined and configured within the system, serve as the unit of measurement for the calculation. The Parking Charge Calculator multiplies these charges by the elapsed time, providing a precise and automated method for determining the parking fee for a given vehicle. This approach not only ensures accuracy in charge computation but also eliminates the need for manual intervention, promoting efficiency and reducing potential errors. The integration of ANPR technology and a sophisticated database system not only expedites the identification process but also contributes to the overall reliability and security of the Parking Charge Calculator. By automating the parking charge calculation based on vehicle entry and exit times, this system offers a seamless and transparent method for both parking lot operators and vehicle owners, enhancing the overall parking experience and fostering a more technologically advanced approach to parking Management

Automatic Recognition Technology of Library Books Based on Convolutional Neural Network Model

Background: The development of technological devices has changed many facets of our lives, particularly the way we engage with information and learning. The advent of automated technology for identification has had a had a revolutionary effect on how we read and organize books within the context of books and data searches. It starts by solving the difficulties in analyzing photos of book pages by using methods such as distortion rectification and book separation. Objective: The research compares the effectiveness of the suggested method with traditional straight-line identification techniques using real-world testing. Methodology: The Skip-Gram model in Word2Vec is used to accurately represent spoken language, allowing word vectors to be generated and input data to be preprocessed for CNN. The results show that the methodology created regarding the present investigation works better than alternatives concerning accuracy and efficiency during line identification. Result:This work advances the field of book suggestion systems by presenting a strong and effective method that leverages CNNs. The findings demonstrate deep learning techniques may be used to optimize system recommendations and improve customer service and happiness in a variety of contexts. This technique creates a bridge between natural language processing and picture evaluation and opens up new possibilities for suggestion advancement along with user satisfaction. Doi: 10.28991/HIJ-2024-05-01-015 Full Text: PDF

Positional relationship between ball and fingers for accurate baseball pitching.

Accurately throwing an object to a target position repeatedly is one of the specific human motor skills. The final arrival position of a thrown ball can be determined by its physical state at release. In baseball pitching, reducing the variability of the velocity angle of the ball at release (release angle) is important for reducing the variability of the pitch location. Although previous studies have suggested that hand and finger movements are important for accurate throwing, their relationship with the release angle has not yet been investigated in detail. This study focused on the positional relationship between the ball and fingers, which is considered to be closely related to ball movement as an action point of the force, and examined its relationship with the variability of release angle. To obtain accurate finger positions relative to the ball without impeding movement or sensation, an automatic image recognition technology based on deep learning was employed. This approach revealed a noteworthy correlation between the lower middle finger positions prior to acceleration peaks and the reduced variability in release angle, emphasizing the importance of consistent finger positioning during the pre-release phase. This finger positioning of the pitchers with low variability in the release angle is suggested to be robust against the spatial variability of ball movement.

A COMPARISON STUDY ON DEEP LEARNING MODELS FOR BUILDING ROOFTOP CLASSIFICATION

Abstract. The availability of semantic information about a cityscape is essential for understanding and analysing urban processes. Automatic gathering of such information is important due to the enormous amount of data. A great number of building features could be gained solely by visual inspections. Therefore, it is meaningful to utilize recent advancements in automatic image recognition technologies to extract these properties automatically.This paper proposes an optimized solution for the classification of rooftops from aerial imagery based on a deep learning model using Convolutional Neural Networks (CNNs). It describes the architecture of the network, the training procedure and important hypermeters. A model analysis using advanced interpretability and explainability tools is conducted. The model’s superiority is demonstrated by comparing its performance against several state-of-the-art image classification architectures, including CNN-based ones such as Xception and Efficientnet, pure Visual Transformers (ViTs) based architectures such as BEiT, and hybrid architectures.

4-degree-of-freedom voice-controlled robotic arm

People with impairments who utilize robotic arms have a great deal more independence in their daily lives. The robotic arm can help persons with specific needs by using voice control, which is practical. The working mechanism is fashioned to have four degrees of freedom (4-DOF) with the inclusion of a jaw or gripper. The Arduino microcontroller, the speech module, and the arm are the main components of the robotic arm, which is a lightweight model driven by four motors. The servo motor is used to supply the rotational motion in the arm’s three rotary joints and end effector. Rotation in two directions is made possible by the geared DC motors employed, which alter their direction of rotation if the polarity is flipped. The automatic voice recognition technology aids in comprehending spoken words picked up by a microphone. The HM2007 processor, the brains of the speech recognizer, is necessary for the speech recognizer to function. Digital commands are created from analog speech input. These commands are used as input to the Arduino thereby resulting in a system whereby the human voice gives a continuous control signal to operate a real and functional robotic arm.

Functional connectivity-enhanced feature-grouped attention network for cross-subject EEG emotion recognition

Electroencephalogram (EEG)-based automatic emotion recognition technologies are gaining significant attention and have become crucial in the field of brain–computer interfaces (BCIs). In particular, deep learning methods have been widely used in emotion recognition in recent years. However, most existing methods generally tend to focus on EEG spatiotemporal information, which ignores the potential relationships between brain activity signals and the differences in functional connectivity under different emotions. Here, we raise a functional connectivity-enhanced feature-grouped attention network (FC-FAN) for cross-subject emotion recognition. The FC-FAN model developed is a dual-input model. One input consists of the differential entropy data derived from the original EEG signals, while the other input comprises the functional connectivity data obtained through the calculation of the phase synchronization index. Then the primary EEG features of the two groups’ input data are extracted through two specific residual blocks. Next, the designed time-series feature grouped attention module (TFGAM) and functional connectivity feature grouped attention model (F2GAM) are utilized to mark interested information or suppress uninterested features for the two groups’ features, respectively. Finally, generated information interacts through a fusion operator. The designed framework could not only sufficiently learn the spatiotemporal features of EEG signals but also clearly analyze nonlinear correlations between electrode signals. Comprehensive tests confirm that the FC-FAN has an excellent effect on subject-independent emotion recognition tasks.

Revolutionizing law enforcement: The role of artificial intelligence in license plate recognition

This paper aims to offer a comprehensive review of the current state of the art in artificial intelligence (AI) as applied to license plate recognition. With the rapidly evolving nature of AI technology, deep learning approaches have gained popularity in license plate recognition, as exemplified by the success of AlphaGo. The diversity of AI in license plate recognition is notable, with numerous studies proposing systems that have achieved high accuracy in segmentation and recognition. The process of reading license plates is complex and involves several stages, including image capture, pre-processing, license plate identification, character segmentation, and recognition. Law enforcement widely uses automatic license plate recognition (ALPR) technology for detecting and preventing criminal activities, tracking stolen vehicles, and identifying suspects. Additionally, ALPR technology can monitor travel time on significant roadways, which can provide the Department of Transportation with useful data for efficient traffic management. Overall, this paper highlights the importance of AI in license plate recognition and its potential to revolutionize the field.

Remote Sensing Classification of Offshore Seaweed Aquaculture Farms on Sample Dataset Amplification and Semantic Segmentation Model

Satellite remote sensing provides an effective technical means for the precise extraction of information on aquacultural areas, which is of great significance in realizing the scientific supervision of the aquaculture industry. Existing optical remote sensing methods for the extraction of aquacultural area information mostly focus on the use of image spatial features and research on classification methods of single aquaculture patterns. Accordingly, the comprehensive utilization of a combination of spectral information and deep learning automatic recognition technology in the feature expression and discriminant extraction of aquaculture areas needs to be further explored. In this study, using Sentinel-2 remote sensing images, a method for the accurate extraction of different algae aquaculture zones combined with spectral information and deep learning technology was proposed for the characteristics of small samples, multidimensions, and complex water components in marine aquacultural areas. First, the feature expression ability of the aquaculture area target was enhanced through the calculation of the normalized difference aquaculture water index (NDAWI). Second, on this basis, the improved deep convolution generative adversarial network (DCGAN) algorithm was used to amplify the samples and create the NDAWI dataset. Finally, three semantic segmentation methods (UNet, DeepLabv3, and SegNet) were used to design models for classifying the algal aquaculture zones based on the sample amplified time series dataset and comprehensively compare the accuracy of the model classifications for achieving accurate extraction of different algal aquaculture information within the seawater aquaculture zones. The results show that the improved DCGAN amplification exhibited a better effect than the generative adversarial networks (GANs) and DCGAN under the indexes of structural similarity (SSIM) and peak signal-to-noise ratio (PSNR). The UNet classification model constructed on the basis of the improved DCGAN-amplified NDAWI dataset achieved better classification results (Lvshunkou: OA = 94.56%, kappa = 0.905; Jinzhou: OA = 94.68%, kappa = 0.913). The algorithmic model in this study provides a new method for the fine classification of marine aquaculture area information under small sample conditions.

Application of Image Sensing System in Mineral/Rock Identification: Sensing Mode and Information Process

Artificial intelligence promotes the development of mineral/rock automatic recognition technology to reduce labor costs and reliance on personal experience. With instrument digitization, image recognition has played an increasingly important role. This review focuses on the image sensing system applied to mineral/rock identification and redivides this system into a sensing module, an imaging module, and an information processing module. The sensing and imaging mechanisms are summarized via technologies of sensing, imaging, and display. The information processing is classified into single‐modal or multimodal recognition. This review analyzes the limitations of single‐modal recognition and discusses the advantages and significance of multimodal recognition. The challenges of current mineral/rock identification are concluded, such as public database, multimodal learning, and portable sensing system. The interdisciplinary sensing technologies are discussed, and application feasibility is explored. This review aims to provide a reference for the research community in automatic mineral/rock identification and sensing system optimization.

The Use of ELSA Speak as a Mobile-Assisted Language Learning (MALL) towards EFL Students Pronunciation

This study analyses how students' pronunciation can be improved through the use of the ELSA Speak as a Mobile-Assisted Language Learning (MALL) tool. ELSA stands for English Language Speech Assistant. The speech recognition function of ELSA Speak is powered by artificial intelligence (AI), and it provides users with feedback to help them improve their pronunciation, which improves their ability to speak English. This study investigates the influence and potential of the ELSA Speak application on the acquisition of pronunciation by employing a qualitative research method and a strategy that relies on library research. The findings highlight the significance of having an excellent content design, an efficient pedagogical and instructional design, gamified learning elements, automatic voice recognition technology, a secure learning environment, and individualised digital feedback. The findings of the study indicate that there is considerably significant potential for the ELSA Speak application to help students improve their pronunciation skills. The consequences of this research emphasize the necessity of utilising cutting-edge mobile applications for language learning and highlight the possible benefits for students interested in improving their pronunciation proficiency.&#x0D; &#x0D; Keywords: Elsa Speak; Mobile-Assisted Language Learning; Students' Pronunciation; Pronunciation Improvement.

Political Visual Literacy

Abstract Visual politics is a fast-growing field and much of it is focused on images that inspire criticism. This tendency results in a lack of attention to oppressive visual practices. A political visual literacy approaches all visual practices as being layered with different “visual truths” that were developed in response to political commitments over time. These “visual truths” inflected visual practices in ways that may outlive the political settings in which they were first articulated. Most important of these is the desire to capture human difference that framed the development of visual technologies and is still embedded in a range of visual practices. The methodology I develop here links a conjunctural analysis of visual tools and practices and the visual truths implicated in them with their operationalisation by actors as cultural tools through the framework of mediated action. I develop this approach by interrogating two layered and harmful visual practices: the White-centrism of visual technologies and the racialised origins of transphobic visualities in automatic gender recognition technology.

E-Challan Generation System using ANPR with Desktop Application

Abstract: Efficiently overseeing adherence to traffic regulations is an arduous task faced by authorities in light of population growth and increased vehicular activity on roads. The conventional approach to managing road users who break laws involves time-intensive manual processes that interfere with smooth transportation operations. This paper proposes a promising tactic for automating the production of E-challans through incorporation of Automatic Number Plate Recognition (ANPR) technology. By installing mounted cameras with ANPR capabilities alongside CCTV equipment which utilizes image processing together with optical character recognition technology (OCR), swift automated identification of infringing drivers can be achieved through read-outs from their vehicle registration plates. Ultimately, this method seeks to lessen reliance on human resources while enhancing the effectiveness of law enforcement activities

Smart Attendace System Using ML

Abstract: The world is changing, and automatic face recognition (AFR) technologies have advanced greatly. Smart Attendance with Real-Time Face Recognition is a practical method for managing the student attendance system on a daily basis. The technique of recognising a student's face for collecting attendance using facial biometrics based on high-definition monitor footage and other information technology is known as a face recognition-based attendance system. In my face recognition project, a computer system will be able to quickly and accurately find and recognise human faces in pictures or videos being taken using a security camera. Many algorithms and methods have been created to enhance facial recognition ability, however deep learning is the idea that will be used in this case. It assists in turning the video frames into photos so that the student's face can be quickly identified for their attendance and the attendance database can be quickly updated automatically.

Classifying Microscopic Images of Reactive Lymphocytosis Using Two-Step Tandem AI Models

The practical applications of automatic recognition and categorization technology for next-generation systems are desired in the clinical laboratory. We approached the identification of reactive lymphocytosis using artificial intelligence (AI) technology and studied its clinical usefulness for blood smear screening. This study created one- and two-step AI models for the identification of reactive lymphocytosis. The ResNet-101 model was applied for deep learning. The original image set for supervised AI training consisted of 5765 typical nucleated blood cell images. The subjects for clinical assessment were 25 healthy cases, 25 erythroblast cases, and 25 reactive lymphocytosis cases. The total accuracy (mean ± standard deviation) of the one- and two-step models were 0.971 ± 0.047 and 0.977 ± 0.024 in healthy, 0.938 ± 0.040 and 0.978 ± 0.018 in erythroblast, and 0.856 ± 0.056 and 0.863 ± 0.069 in reactive lymphocytosis cases, respectively. The two-step AI model showed a sensitivity of 0.960 and a specificity of 1.000 between healthy and reactive lymphocytosis cases. As our two-step tandem AI model showed high performance for identifying reactive lymphocytosis in blood smear screening, we plan to apply this method to the development of AI models to differentiate reactive and neoplastic lymphocytosis.

Real Time Web-Based Smart Attendance System using AI

The performance of automatic face recognition (AFR) technology has improved in recent years. As a result, these systems are frequently employed for security and business purposes. a face recognition programme that runs automatically in the background while a university tracks employee attendance. Hence, Smart Attendance with Real Time Face Recognition might be a global solution for managing personnel on a daily basis. The work is incredibly challenging since the critical temporal background subtraction during an image remains a challenge. Real-time face detection is done using easy, quick Principal Component Analysis, which has a high rate of success in identifying the faces found. As a result, we have developed a module in which an employee's attendance will be recorded automatically by recognising their face using face recognition technology. Their face will be saved in the database after this one-time enrollment process. Since enrolling a face is a one-time activity, we need a mechanism. Every employee will have a unique employee ID that is their own roll number. Every employee's presence will be updated within a database. Results of the proposed system demonstrated that it performs better than the manual attendance system. The marking of attendance follows employee identification. This application provides the customer with considerably more solutions and accurate leads than the conventional attendance and leave.

Mobile robot: automatic speech recognition application for automation and STEM education

Nowadays, robots are widely applied in life as well as industrial production, medicine, rescue, learning, and entertainment. There are many kinds of robots using different modern technologies such as manipulators, movement, and biomimetic robots; each type is applied according to its own role. Robots are also applied in training, especially in STEM training implemented with mobile robots. In this paper, we proposed models to design mobile robot combined with Android OS Tablet with a user interaction screen, voice control model, and AIDL IPC interactive model for remote control. We research the mobile robot using RockChip AI Processor RK3399Pro, which focuses on understanding the structure and operating mechanism of the AI Processor that RockChip robot uses, along with researching Android with IPC AIDL architecture and Automatic Voice Recognition technology to control robot to turn left, turn right, move forward, backward, and stop the movement by voice with three languages: Korean, English, and Vietnamese. We experimented on 1850 recording audio files including 1013 female voices and 837 male voices with ‘latest_short,’ ‘latest_long,’ and ‘command_and_search’ models. The results obtained a fairly high average confidence of 95% for English and Vietnamese languages and 87% for Korean. From results of this study, it will serve as a foundation for the research and development of a system to support students in learning by interacting with mobile robots through voice with multi-language. STEM education in the form of gamification of the curriculum, especially for extensive research on training programs can be applied in mobile robots with more flexibility.

Improving Modulation Recognition Using Time Series Data Augmentation via a Spatiotemporal Multi-Channel Framework

Automatic modulation recognition technology with deep learning has a broad prospective owing to big data and computing power. However, the accuracy of modulation recognition largely depends on the massive volume of data and the applicability of the model. Here, to eliminate the difficulties of manual feature extraction, a low accuracy, and a small sample dataset, we propose an effective recognition method that combines time series data augmentation with a spatiotemporal multi-channel learning framework. Compared with other advanced network models, the results showed that the method gave a positive index in the order of 93.5% for ten modulation signal types, which was increased by at least 15%. Especially for QAM16 and QAM64 signals, the average recognition accuracy was improved by nearly 50% at SNRs as low as −2 dB, showing a significant recognition performance. The proposed method provides an attractive method for signal modulation recognition in wireless or wired communication fields.

Transferable Adversarial Attacks against Automatic Modulation Classifier in Wireless Communications

Deep neural network-based automatic modulation recognition (AMR) technology has become an increasingly important area due to the advantages of self-extraction of features and high identification accuracy. Based on the view of security threats to machine learning classifiers, we investigate the influence of adversarial samples on the AMR model in this paper. The traditional method is based on label gradient attack without taking advantage of the feature-level transferability, resulting in the attack effect that is not perfect. So, we exploit the feature-level transferability property that could be met to fulfill realistic imperceptibility and transfer needs. In this paper, firstly, we proposed an AMR scheme with high recognition accuracy as our attack model. Secondly, we proposed a transferable attack method based on a feature gradient-based, which increases perturbation to clean signal based on features space. Finally, we introduce a new attack strategy, in which we select two original and one adversarial target signal sample as the input of triplet loss to achieve higher attack strength and high transferability. Meanwhile, this paper proposes indicators of signal characteristics to test the effectiveness of our proposed attack method. Based on experimental results, our proposed feature gradient-based adversarial attack method outperforms the currently labeled gradient attack methods regarding attack effectiveness and transferability.