Recognition Algorithm Research Articles

Application and Analysis of the MFF-YOLOv7 Model in Underwater Sonar Image Target Detection

The need for precise identification of underwater sonar image targets is growing in areas such as marine resource exploitation, subsea construction, and ocean ecosystem surveillance. Nevertheless, conventional image recognition algorithms encounter several obstacles, including intricate underwater settings, poor-quality sonar image data, and limited sample quantities, which hinder accurate identification. This study seeks to improve underwater sonar image target recognition capabilities by employing deep learning techniques and developing the Multi-Gradient Feature Fusion YOLOv7 model (MFF-YOLOv7) to address these challenges. This model incorporates the Multi-Scale Information Fusion Module (MIFM) as a replacement for YOLOv7’s SPPCSPC, substitutes the Conv of CBS following ELAN with RFAConv, and integrates the SCSA mechanism at three junctions where the backbone links to the head, enhancing target recognition accuracy. Trials were conducted using datasets like URPC, SCTD, and UATD, encompassing comparative studies of attention mechanisms, ablation tests, and evaluations against other leading algorithms. The findings indicate that the MFF-YOLOv7 model substantially surpasses other models across various metrics, demonstrates superior underwater target detection capabilities, exhibits enhanced generalization potential, and offers a more dependable and precise solution for underwater target identification.

An Environment Recognition Algorithm for Staircase Climbing Robots

For deformed wheel-based staircase-climbing robots, the accuracy of staircase step geometry perception and scene mapping are critical factors in determining whether the robot can successfully ascend the stairs and continue its task. Currently, while there are LiDAR-based algorithms that focus either on step geometry detection or scene mapping, few comprehensive algorithms exist that address both step geometry perception and scene mapping for staircases. Moreover, significant errors in step geometry estimation and low mapping accuracy can hinder the ability of deformed wheel-based mobile robots to climb stairs, negatively impacting the efficiency and success rate of task execution. To solve the above problems, we propose an effective LiDAR-Inertial-based point cloud detection method for staircases. Firstly, we preprocess the staircase point cloud, mainly using the Statistical Outlier Removal algorithm to effectively remove the outliers in the staircase scene and combine the vertical angular resolution and spatial geometric relationship of LiDAR to realize the ground segmentation in the staircase scene. Then, we perform post-processing based on the point cloud map obtained from LiDAR SLAM, extract the staircase point cloud and project and fit the staircase point cloud by Ceres optimizer, and solve the dimensional information such as depth and height of the staircase by combining with the mean filtering method. Finally, we fully validate the effectiveness of the method proposed in this paper by conducting multiple sets of SLAM and size detection experiments in real different staircase scenarios.

Band Correlation‐Based Multichannel Multiscale Convolution Network for Intelligent Interference Recognition

AbstractIn recent years, with the development and extensive application of wireless communication technology, the communication system should have stronger anti‐Jamming ability. Therefore, interference recognition is particularly important as a prerequisite for anti‐interference. However, the existing traditional and intelligent interference recognition algorithms have problems such as complicated feature extraction and low recognition accuracy under low interference‐to‐noise ratio. In order to solve the above problems, this paper introduces parallel multi‐channel multi‐scale convolution to improve the speed and accuracy of network recognition. In addition, combined with frequency band correlation and long‐short‐term memory network (LSTM), an innovative wireless communication interference identification model based on frequency band correlation is proposed, which uses LSTM to detect the frequency band correlation of interference signals and improve the accuracy of interference identification under low Jamming noise ratio (JNR). Experiments prove that the model proposed in this article has faster recognition speed and better generalization. The introduction of frequency band correlation increases the recognition accuracy to more than 99% with low JNR. Therefore, the model proposed in this paper is an effective and available model in complex electromagnetic environments. © 2024 Institute of Electrical Engineers of Japan and Wiley Periodicals LLC.

3D Non-separable Moment Invariants and Their Use in Neural Networks

Recognition of 3D objects is an important task in many bio-medical and industrial applications. The recognition algorithms should work regardless of a particular orientation of the object in the space. In this paper, we introduce new 3D rotation moment invariants, which are composed of non-separable Appell moments. We show that non-separable moments may outperform the separable ones in terms of recognition power and robustness thanks to a better distribution of their zero surfaces over the image space. We test the numerical properties and discrimination power of the proposed invariants on three real datasets—MRI images of human brain, 3D scans of statues, and confocal microscope images of worms. We show the robustness to resampling errors improved more than twice and the recognition rate increased by 2–10 % comparing to most common descriptors. In the last section, we show how these invariants can be used in state-of-the-art neural networks for image recognition. The proposed H-NeXtA architecture improved the recognition rate by 2–5 % over the current networks.

Comparative Analysis of Face Recognition Algorithms for Facial Recognition in Diverse Environments

Facial recognition technology has evolved significantly over the last five decades and plays a central role in various applications such as biometrics, information security, access control, law enforcement and surveillance. In this study, the performance of two face recognition algorithms, Dlib and FaceNet, is evaluated using datasets obtained from video recordings in different environments. The Dlib algorithm uses the Histogram of Oriented Gradients (HOG) method for face detection, while FaceNet uses the Multi-Task Cascaded Convolutional Neural Network (MTCNN). The experimental results show that both algorithms achieve high accuracy in controlled environments, with Dlib showing greater robustness in complex scenarios. This study makes an important contribution to this topic by presenting a comparative analysis of the face recognition performance of the OpenFace, ArcFace, Exadel, and Dlib methods under different environmental conditions and scenarios. The results show that while the tested methods achieve high accuracy in controlled environments, their performance differs in more com-plex environments.In the results, OpenFace and ArcFace showed lower success rates than the other two algorithms. In particu-lar, Dlib proved superior in dynamic and challenging scenarios, achieving an overall accuracy of 96.1% compared to 94.6% for Exadel. Exadel, on the other hand, performed slightly better in certain controlled environments, highlighting its potential strength in certain applications. These results emphasize the importance of selecting the appropriate algorithm based on the specific environmental conditions and requirements of the application. This research not only improves our understanding of the performance characteristics of leading facial recognition technologies, but also provides practical insights into their use in real-world applications.

Vision-based weld detection and localization for climbing robots in large-scale storage tank inspections

Abstract Climbing robots are considered an effective solution for inspecting welds on the walls of large storage tanks. For these robotic systems, the efficient and accurate identification and localizing of weld seams are crucial prerequisites for ensuring precise weld seam tracking. In this paper, we investigate machine vision-based algorithms for feature recognition and localization of weld seams on tank walls for inspection of weld seams by a climbing robot. First, we designed the model of the image algorithm to extract the weld features of the tank walls. After extracting the weld features, we propose the novel idea of feature discretization and a Min-outer Rectangle Fitting Algorithm (MRFA), which will achieve the fitting of rectangular features on the discretized weld features. We constructed a mathematical model for calculating the orientation angle of the rectangular box based on the extracted rectangular boxes. This model allows for the real-time and efficient extraction of the rectangular feature’s pose information (x, y, θ). We also propose an efficient method for calculating the curvature of a curve trajectory. The experimental results demonstrate that the proposed image algorithm model and MRFA effectively identify weld features on the storage tank wall surface, while simultaneously achieving high-accuracy feature localization. Positioning errors are maintained within 3 mm for position and 3 degrees for azimuth, indicating both high precision and robustness. Additionally, the algorithm processes each image in approximately 80 milliseconds. The lightweight and efficient design of the proposed model allows it to be easily deployed on a climbing robot for weld seam detection and tracking on tank walls.

Information Technology for Sound Analysis and Recognition in the Metropolis based on Machine Learning Methods

The goal of designing and implementing an intelligent information system for the recognition and classification of sound signals is to create an effective solution at the software level, which would allow analysis, recognition, classification and forecasting of sound signals in megacities and smart cities using machine learning methods. This system can help people in various fields to simplify their lives, for example, it can help farmers protect their crops from animals, in the military it can help with the identification of weapons and the search for flying objects, such as drones or missiles, in the future there is a possibility for recognizing the distance to sound, also, in cities can help with security, so a preventive response system can be built, which can check if everything is in order based on sounds. Also, it can make life easier for people with impaired hearing to detect danger in everyday life. In the part of the comparison of analogues of the developed product, 4 analogues were found: Shazam, sound recognition from Apple, Vocapia, and SoundHound. A table of comparisons was made for these analogues and the product under development. Also, after comparing analogues, a table for evaluating the effects of the development was built. During the system analysis section, a variety of audio research materials were developed to indicate the characteristics that can be used for this design: period, amplitude, and frequency, and, as an example, an article on real-world audio applications is shown. A precedent scenario is described using the RUP methodology and UML diagrams are constructed: Diagram of use cases; Class diagram; Activity chart; Sequence diagram; Diagram of components; and Deployment diagram. Also, sound data analysis was performed, sound data was visualized as spectrograms and sound waves, which clearly show that the data are different, so it is possible to classify them using machine learning methods. An experimental selection of the machine learning method as staandart clasificers for building a sound recognition model was made. The best method turned out to be SVC, the accuracy of which reflects more than 30 per cent. A neural network was also implemented to improve the obtained results. The result of training a model based on a neural network during 100 epochs achieved a result of 97.7% accuracy for training data and 47.8% accuracy when checking performance on test data. This result should be higher, so it is necessary to consider improving recognition algorithms, increasing the amount of data, and changing the recognition method. Testing of the project was carried out, showing its operation and pointing out shortcomings that need to be corrected in the future.

High-resolution image processing and entity recognition algorithm based on artificial intelligence

Abstract Objective With the popularity of high-resolution devices such as high-definition, ultra-high-definition televisions, and smartphones, the demand for high-resolution images is also increasing, which puts forward higher requirements for high-resolution image processing and entity recognition technology. Method This article introduced the research progress and application of high-resolution image processing and entity recognition algorithms from the perspective of artificial intelligence (AI). First, the important role of AI in high-resolution image processing and entity recognition was introduced, and then the applications of deep learning-based algorithms in high-resolution image grayscale equalization, denoising, and deblurring were introduced. Subsequently, the application of AI-based object detection and image segmentation algorithms in entity recognition was explored, and the superiority of AI-based high-resolution image processing and entity recognition algorithms was verified through training and testing. The accuracy of the model was verified through testing experiments. Finally, a summary and outlook were made on high-resolution image processing and entity recognition algorithms based on AI. Result After experimental testing, it was found that high-resolution image processing and entity recognition based on AI had higher efficiency, and the overall image recognition ability was improved by 29.6% compared to traditional image recognition models. The recognition speed and accuracy were also improved. Conclusion High-resolution image processing and element recognition algorithms based on AI enabled observers to see the detailed information in the image more clearly, thus improving the efficiency and accuracy of image analysis. Through continuous improvement of algorithm performance, real-time application, and expansion of cross-disciplinary applications, people can look forward to the development of more advanced and powerful image processing and entity recognition technologies, which will bring huge impetus to research and application in various fields.

Research on Minority Character Recognition - Taking Small Seal Font as an Example

Text recognition is one of the important fields of computer vision and is widely used in automated office, assisted reading and other fields. With the continuous development of deep learning, the visual language text recognition method generated by the combination of optical character recognition and natural language processing involves the technology of recognizing text information from images or videos, which greatly improves the machine's understanding ability and interaction efficiency. However, for some niche fonts, such as Urdu, Xixia, and Xiaozhuan, these characters have complex structures, diverse strokes, and sloppy writing, making them a challenging field in text recognition. Taking Small Seal Scripts, also called Xiaozhuan, as an example, this paper summarizes the advantages and disadvantages of the current Xiaozhuan text recognition algorithm, and proposes algorithm improvement suggestions for the segmentation of Xiaozhuan characters on seals. Xiaozhuan fonts are quite different from modern Chinese characters, mainly reflected in the stroke structure, writing style, and the correlation between characters. It has smoother lines and many glyphs are unique. This complexity makes it difficult to directly apply traditional Optical Character Recognition (OCR) technology. Therefore, many studies have combined image preprocessing, character segmentation, and deep learning models for automatic recognition. Future research should focus more on the generalization and lightweight of the model so that it can run on devices with limited computing resources.

Research on Emotion Recognition Algorithm of Piano Music Based on Multimodal Learning

The purpose of this study is to propose an emotion recognition algorithm for piano music based on multimodal learning. Multi-modal learning is a learning method that combines various types of data. By combining different modal information, learning tasks can be completed more comprehensively and accurately. In the field of music emotion recognition, multimodal learning can extract the emotional features of music from different angles, and improve the recognition accuracy and robustness. In this study, the multi-modal learning method is mainly used, and the representative features are extracted by combining the audio signal, score information and the performance of the piano music, and the machine learning algorithm is used to classify the emotions. The audio signal is preprocessed, including noise reduction, cutting and other operations, and then the features of the audio signal are obtained by feature extraction methods, such as short-time Fourier transform (STFT) and formant analysis (RPA). Digitally convert the score, extract the structure, melody, harmony and other characteristics of the music, and obtain the emotional information of the music by analyzing the relationship between symbols and notes in the score. By analyzing the performance skills, strength, speed and other parameters of the player, the characteristics related to the emotional expression of the player are extracted. Machine learning algorithm is used to train and classify the fusion features, and the emotion recognition of piano music is realized. In this study, the emotion recognition algorithm of piano music based on multimodal learning can mine the emotion information of music at different levels, improve the accuracy and robustness of emotion recognition, and provide useful reference for the field of music emotion calculation.

Research progress and prospect of key technologies of fruit target recognition for robotic fruit picking

It is crucial for robotic picking fruit to recognize fruit accurately in orchards, this paper reviews the applications and research results of target recognition in orchard fruit picking by using machine vision and emphasizes two methods of fruit recognition: the traditional digital image processing method and the target recognition method based on deep learning. Here, we outline the research achievements and progress of traditional digital image processing methods by the researchers aiming at different disturbance factors in orchards and summarize the shortcomings of traditional digital image processing methods. Then, we focus on the relevant contents of fruit target recognition methods based on deep learning, including the target recognition process, the preparation and classification of the dataset, and the research results of target recognition algorithms in classification, detection, segmentation, and compression acceleration of target recognition network models. Additionally, we summarize the shortcomings of current orchard fruit target recognition tasks from the perspectives of datasets, model applicability, universality of application scenarios, difficulty of recognition tasks, and stability of various algorithms, and look forward to the future development of orchard fruit target recognition.

A positioning method for AUV based on guiding light source

This paper proposes a method for determining the position of underwater vehicles using light vision guidance. In order to solve the problem of limited perception distance and unclear underwater images, the illumination device is installed on the recovery platform as a light source to enhance the feature information of the recovery platform. The collected underwater light source images are enhanced by the dark channel prior defogging algorithm, and then the YOLOv5 algorithm is used to recognize the targets. Through experiments, it is verified that the image enhancement algorithm and target recognition algorithm are effective. Using the recognition results, the positioning accuracy of the binocular camera ranging algorithm is verified, and the results indicate that the proposed method can accurately determine the distance of the AUV relative to the recovery platform.

Application of Pattern Recognition Methods to Study Spatial Localization of Polymetallic Mineralization in the Altai–Sayan Region

The Altai–Sayan mountain-folded belt is analyzed with the purpose of (1) revealing peculiarities of localization of large-scale polymetallic mineralization in the lineament-block structure of the region and (2) determining the geophysical and geomorphic peculiarities of the locations of these deposits using the Cora-3 pattern recognition algorithm. The lineament-block structure of the region is determined using morphostructural zoning. A spatial correlation between large and superlarge polymetallic deposits and morphostructural nodes is revealed. Based on this correlation, a dichotomy problem is solved, which is to divide the entire set of nodes in the region into two classes—ore-bearing and non-ore bearing. For this purpose, we used the Cora-3 logical recognition algorithm with training, for which the input data are geomorphological and geophysical parameters of the nodes. The training set of the algorithm was composed of the nodes where large and superlarge polymetal deposits are known. At the training stage, the algorithm identified the sets of the characteristic features that are peculiar to each class. Based on these features, all the nodes in the region were divided into ore-bearing and non-ore-bearing ones. As a result of recognition, the nodes in which deposits of the considered types and sizes are known were classified as ore-bearing, and, in addition to them, another 11 nodes were identified that meet the features determined in the work and can be considered potentially ore-bearing.

An integrated microfluidic device for sorting of tumor organoids using image recognition.

Tumor organoids present a challenge in drug screening due to their considerable heterogeneity in morphology and size. To address this issue, we proposed a portable microfluidic device that employs image processing algorithms for specific target organoid recognition and microvalve-controlled deflection for sorting and collection. This morphology-activated organoid sorting system offers numerous advantages, such as automated classification, portability, low cost, label-free sample preparation, and gentle handling of organoids. We conducted classification experiments using polystyrene beads, F9 tumoroids and patient-derived tumor organoids, achieving organoid separation efficiency exceeding 88%, purity surpassing 91%, viability exceeding 97% and classification throughput of 800 per hour, thereby meeting the demands of clinical organoid medicine.

Flexible Photoacoustic Device Integrating Electroluminescence, Piezoelectric Vibration, and Pressure Sensing.

Marine organisms rely on the integration of light, sound, and pressure sensing for essential activities such as communication, hunting, and predator evasion. Current bioinspired devices, however, are limited by independent sensing modules, leading to complex designs and reduced adaptability. This study addresses these challenges by developing a multifunctional flexible photoacoustic device (MFPAD) that integrates electroluminescence, piezoelectric vibration, and pressure sensing into a single structure. Using airflow-assisted electrospinning, the device's fibrous layers were fabricated to optimize mechanical properties and enable synchronized responses to luminescence, sound, and pressure. Evaluations showed that MFPAD is highly sensitive to pressure changes, with corresponding light emissions, and efficient sound output modulated by input voltage and frequency. The device also enabled real-time posture tracking in robotic fish using an image recognition algorithm based on luminescence. Additionally, deep learning models allowed for the reconstruction of missing key points during motion. These findings demonstrate the promising potential of MFPAD for advancing bioinspired systems with multimodal sensing capabilities, offering a scalable solution for complex environments.

An Automatic Modulation Recognition Algorithm Based on Time–Frequency Features and Deep Learning with Fading Channels

Automatic modulation recognition (AMR) stands as a crucial core technology within the realm of signal processing and perception, playing a significant part in harsh electromagnetic environments. The time–frequency image (TFI) of communication signals can manifest modulation characteristics and serve as a foundation for signal modulation recognition and classification. However, under the influence of the electromagnetic environment, communication signals are exposed to varying degrees of interference, which poses a challenge to the recognition of modulation types. Taking into account the effects of interference and channel fading, this paper introduces a communication signal modulation recognition algorithm based on deep learning (DL) and time–frequency analysis. This approach employs short-time Fourier transform (STFT) to generate time–frequency diagrams from time-domain signals. Subsequently, it binarizes the image and feeds it as input data to the neural network. Our research presents a composite deep convolutional neural network (CNN) architecture known as the composite dense-residual neural network (CDRNN). This architecture focuses on enhancing the feature extraction and identification, aiming to achieve accurate recognition of modulation types in harsh electromagnetic environments. Finally, simulation results validate that the proposed deep learning algorithm holds remarkable advantages in boosting the accuracy of modulation type recognition with better adaptability. The algorithm shows better performance even in harsh electromagnetic environments. When the signal-to-noise ratio (SNR) is 18 dB, the recognition accuracy can reach 92.1%.

Model-agnostic meta-learning for EEG-based inter-subject emotion recognition.


Developing an efficient and generalizable method for inter-subject emotion recognition from neural signals is an emerging and challenging problem in affective computing. In particular, human subjects usually have heterogeneous neural signal characteristics and variable emotional activities that challenge the existing recognition algorithms from achieving high inter-subject emotion recognition accuracy. 
Approach. 
In this work, we propose a model-agnostic meta-learning algorithm to learn an adaptable and generalizable Electroencephalogram (EEG)-based emotion decoder at the subject's population level. Different from many prior end-to-end emotion recognition algorithms, our learning algorithms include a pre-training step and an adaptation step. Specifically, our meta-decoder first learns on diverse known subjects and then further adapts it to unknown subjects with one-shot adaptation. More importantly, our algorithm is compatible with a variety of mainstream machine learning decoders for emotion recognition.
Main results.
We evaluate the adapted decoders obtained by our proposed algorithm on three Emotion-EEG datasets: SEED, DEAP, and DREAMER. Our comprehensive experimental results show that the adapted meta-emotion decoder achieves state-of-the-art inter-subject emotion recognition accuracy and outperforms the classical supervised learning baseline across different decoder architectures. 
Significance. 
Our results hold promise to incorporate the proposed meta-learning emotion recognition algorithm to effectively improve the inter-subject generalizability in designing future affective brain-computer interfaces (BCIs).

A Research on Two-Stage Facial Occlusion Recognition Algorithm based on CNN

In recent years, pattern recognition has garnered widespread attention, especially in the domain of face recognition. Traditional face recognition methods have certain limitations in unconstrained environments due to factors such as lighting, facial expressions, and poses. Deep learning can be used to address these challenges. This paper proposes a comprehensive approach to face occlusion recognition based on a two-stage Convolutional Neural Network (CNN). Face verification aims at verifying whether two face images belong to the same individual, and it is a more fundamental task compared to face recognition. The process of face recognition essentially involves multiple instances of face verification, sequentially validating different individuals to ultimately determine the corresponding individual for each face. The primary steps in this research include facial detection, image preprocessing, facial landmark localization, facial landmark extraction, feature matching recognition, and 2D image-assisted 3D face reconstruction. A novel two-stage CNN was designed for facial detection and alignment. The first stage of the network is dedicated to the search for facial windows and regressing vector boundaries. The second stage utilizes 2D images to assist in 3D face reconstruction and perform secondary recognition for cases not identified in the first stage. This method demonstrated excellent performance in handling facial occlusions, achieving high accuracy on datasets such as AFW and FDDB. On the test dataset, face recognition accuracy reached 97.3%, surpassing the original network accuracy of 89.1%. This method outperforms traditional algorithms and general CNN approaches. This study achieved efficient face validation and further handling of unrecognized situations, contributing to the enhancement of face recognition system performance.

Evaluation of an automated vessel recognition system: case study in Port of Cristo

Abstract An automated vessel recognition system is a sophisticated technological solution for the automatic detection, identification, and classification of vessels in a maritime environment. This paper presents the evaluation of an automatic vessel recognition system implemented in the Port of Cristo on the island of Mallorca in Spain. The system employs PTZ camera and algorithm for image analysis and object recognition to accurately detect, classify and track different types of small vessels entering or departing from the port. The results revealed that the system successfully detected the event of the arrival and departure of vessels in 95% of the cases. In addition, the system was able to identify the vessel by recognizing the registration number in 75.81% of the cases and classifying the vessel type in 89.07% of the cases. The paper presents the SWOT analysis based on the knowledge and experience gained during the testing period.

Biomechanical sensor signal analysis based on machine learning for human gait classification

Abstract The present study investigates the effect of wearable sensor placements and the use of various machine learning (ML) algorithms for human gait pattern recognition based on temporal gait speeds using wearable multichannel sensor data. Therefore, classifying human gait from features extracted from biomechanical sensor signals and evaluating the effect of using these sensors on gait biomechanics can be successfully achieved with a machine learning approach. In this study, firstly, IMU (Inertial Measurement Unit) and GON (Goniometer) sensor features were extracted for machine learning input using the sliding windows method, and these features were applied to ML classifiers to classify human gait speed. Our experimental findings show that multi- and fusion sensor models provide superior classification performance compared to single sensor models, and the highest accuracy is achieved with the ‘FUS09’ fusion sensor model and SVM-based classifier. The classification accuracies of the FUS10 fusion sensor model, where all sensor parameters were combined, the FUS09 model, where the GON_ANK and IMU_Ft parameters in this model were excluded, are 0.895 and 0.901, respectively. Consequently, wearable biomechanical sensor data and machine learning approach can be easily preferred in multiple human activity recognition.