Intelligent Image Analysis Research Articles

REAL-TIME MOTION IMAGE POSE DECOMPOSITION, CLASSIFICATION AND ANALYSIS

Exercise can boost metabolism and make the body healthier. It also increases metabolic rate, helping to consume more calories and burn fat. Regular exercise can stimulate the brain to secrete endorphins, making people feel relaxed and happy, and improves self-confidence, and has been shown to reduce symptoms in people with depression and anxiety. However, incorrect exercise posture may cause harm to the body, such as torn ligaments or muscle strains, so good exercise posture is needed to improve sports performance. This article proposes to use artificial intelligence image analysis method to decompose fitness exercise posture images and establish exercise posture cycle samples to assist in completing fitness exercises.

Initial Fire Detection Method for Intelligent CCTV Using Deep Learning Technology

Pre-prediction and early detection are important for the prevention of fires. Hence, the integration of intelligent CCTV image analysis technology, deep learning, and artificial intelligence is actively being studied. Intelligent CCTV can analyze visible light images of fires, but the accuracies of the detection targets are low. To solve this problem, a dual sensing experimental device was constructed by combining a visible light sensor and a thermal image sensor, and an algorithm was implemented to predict fire events. An algorithm was constructed to detect three classes (Fire, Smoke, and Person) through RGB sensors and thermal imaging sensors. Additionally, a CNN deep learning method and the YOLOv5 method were used to build a classification neural network. The dual sensing device was able to act as a trigger to predict and detect a fire based on the size of the flame and changes in temperature.

Concrete structure assembly technology based on 3D intelligent image analysis

AbstractAt present, there is no standard practice for temporary connection of component installation. How to make temporary connections to components more efficient and accurate is the key to the construction of cast‐in‐place connection parts in concrete prefabricated buildings. In order to improve the assembly technology of concrete structures, this paper combines three‐dimensional intelligent image analysis technology for simulation, analyzes and compares several common electronic image stabilization algorithms, discusses the video image stabilization technology based on gray projection method in detail, and gives experimental results. Moreover, this paper analyzes and compares common moving target tracking algorithms, such as feature‐based tracking, region matching‐based tracking, dynamic contour‐based tracking, and 3D model‐based tracking. In addition, this paper studies the target tracking algorithm based on the Camshift algorithm, and constructs the assembly model of the intelligent concrete structure with the support of the algorithm. Through experimental verification, it is known that the performance distribution of the concrete structure assembly model based on 3D intelligent image analysis in experimental evaluation is between [81, 89], the experimental study shows that the concrete structure assembly model based on 3D intelligent image analysis can effectively improve the assembly effect of concrete structure.

Application of intelligent tongue image analysis in Conjunction with microbiomes in the diagnosis of MAFLD

BackgroundMetabolic associated fatty liver disease (MAFLD) is a widespread liver disease that can lead to liver fibrosis and cirrhosis. Therefore, it is essential to develop early diagnosic and screening methods. MethodsWe performed a cross-sectional observational study. In this study, based on data from 92 patients with MAFLD and 74 healthy individuals, we observed the characteristics of tongue images, tongue coating and intestinal flora. A generative adversarial network was used to extract tongue image features, and 16S rRNA sequencing was performed using the tongue coating and intestinal flora. We then applied tongue image analysis technology combined with microbiome technology to obtain an MAFLD early screening model with higher accuracy. In addition, we compared different modelling methods, including Extreme Gradient Boosting (XGBoost), random forest, neural networks(MLP), stochastic gradient descent(SGD), and support vector machine(SVM). ResultsThe results show that tongue-coating Streptococcus and Rothia, intestinal Blautia, and Streptococcus are potential biomarkers for MAFLD. The diagnostic model jointly incorporating tongue image features, basic information (gender, age, BMI), and tongue coating marker flora (Streptococcus, Rothia), can have an accuracy of 96.39%, higher than the accuracy value except for bacteria. ConclusionCombining computer-intelligent tongue diagnosis with microbiome technology enhances MAFLD diagnostic accuracy and provides a convenient early screening reference.

Use of an AI Score Combining Cancer Signs, Masking, and Risk to Select Patients for Supplemental Breast Cancer Screening.

Background Mammographic density measurements are used to identify patients who should undergo supplemental imaging for breast cancer detection, but artificial intelligence (AI) image analysis may be more effective. Purpose To assess whether AISmartDensity-an AI-based score integrating cancer signs, masking, and risk-surpasses measurements of mammographic density in identifying patients for supplemental breast imaging after a negative screening mammogram. Materials and Methods This retrospective study included randomly selected individuals who underwent screening mammography at Karolinska University Hospital between January 2008 and December 2015. The models in AISmartDensity were trained and validated using nonoverlapping data. The ability of AISmartDensity to identify future cancer in patients with a negative screening mammogram was evaluated and compared with that of mammographic density models. Sensitivity and positive predictive value (PPV) were calculated for the top 8% of scores, mimicking the proportion of patients in the Breast Imaging Reporting and Data System "extremely dense" category. Model performance was evaluated using area under the receiver operating characteristic curve (AUC) and was compared using the DeLong test. Results The study population included 65 325 examinations (median patient age, 53 years [IQR, 47-62 years])-64 870 examinations in healthy patients and 455 examinations in patients with breast cancer diagnosed within 3 years of a negative screening mammogram. The AUC for detecting subsequent cancers was 0.72 and 0.61 (P < .001) for AISmartDensity and the best-performing density model (age-adjusted dense area), respectively. For examinations with scores in the top 8%, AISmartDensity identified 152 of 455 (33%) future cancers with a PPV of 2.91%, whereas the best-performing density model (age-adjusted dense area) identified 57 of 455 (13%) future cancers with a PPV of 1.09% (P < .001). AISmartDensity identified 32% (41 of 130) and 34% (111 of 325) of interval and next-round screen-detected cancers, whereas the best-performing density model (dense area) identified 16% (21 of 130) and 9% (30 of 325), respectively. Conclusion AISmartDensity, integrating cancer signs, masking, and risk, outperformed traditional density models in identifying patients for supplemental imaging after a negative screening mammogram. © RSNA, 2024 Supplemental material is available for this article. See also the editorial by Kim and Chang in this issue.

Development of a Medical Information System with Data Storage and Intelligent Image Analysis

The relevance of the development of medical information systems for otolaryngology with artificial intelligence is substantiated, the process of developing an effective storage of medical data, taking into account their heterogeneity, is described. The task of storing medical data is complicated by the legal significance, large volume, heterogeneous and complex structure of information. A test version of an automated workstation for medical information systems for storing and processing medical research results has been developed. In the process of creating a new medical information system, the tasks of the first stage were solved: the structure and organization of the medical data warehouse were developed; procedures have been implemented for downloading, storing and viewing the results of medical research in DICOM format, identifying each patient to provide all related research results; criteria for dividing images into categories (TRG, RCT) were developed and implemented. The new system provides support for various types of search and access to the server from any computer in the corporate network of the otolaryngology department. Conditions have been prepared for the implementation of the second stage – integration of an automated workstation of medical information systems into the existing technological process in a medical institution and detection of pathology using an artificial neural network.

Intelligent color image analysis of sintered ores for simple and rapid determination of Fe3O4 concentration

A simple determination of the Fe3O4 concentrations of sintered ores using color images of the samples has been explored. Sintered ore is mainly composed of Fe2O3 (red), Fe3O4 (black), and other white inorganic oxides, so the color of sintered ore could be representative of the relative abundance of the constituents. Two important challenges were addressed to achieve reliable quantitative color image analysis. First, minute dents and bumps (embosses) exist on the sample surface due to inconsistent particle sizes and particle agglomeration, thereby generating dark shadows. Second, small white spots corresponding to inorganic oxide particles were spread throughout acquired images. The white spots yield very high RGB values, which would hamper the translation of the real sample color originating from the iron oxides. Therefore, the segmentations of particle agglomeration-induced shadows and white spots in the sample images were separately executed using Otsu's method and modified fuzzy C-means (MFCM), respectively. Then, color moments and derived variables from the segmented images were employed to determine Fe3O4 concentrations (6.5–10.5 wt%) using extreme gradient boosting (XGBoost). The predicted concentrations from the color analysis correlated well with reference concentrations determined using conventional titration, with a root mean square error of prediction (RMSEP) of 0.39 wt%.

Cytopathology image analysis method based on high-resolution medical representation learning in medical decision-making system

Artificial intelligence has made substantial progress in many medical application scenarios. The quantity and complexity of pathology images are enormous, but conventional visual screening techniques are labor-intensive, time-consuming, and subject to some degree of subjectivity. Complex pathological data can be converted into mineable image features using artificial intelligence image analysis technology, enabling medical professionals to quickly and quantitatively identify regions of interest and extract information about cellular tissue. In this study, we designed a medical information assistance system for segmenting pathology images and quantifying statistical results, including data enhancement, cell nucleus segmentation, model tumor, and quantitative analysis. In cell nucleus segmentation, to address the problem of uneven healthcare resources, we designed a high-precision teacher model (HRMED_T) and a lightweight student model (HRMED_S). The HRMED_T model is based on visual Transformer and high-resolution representation learning. It achieves accurate segmentation by parallel low-resolution convolution and high-scaled image iterative fusion, while also maintaining the high-resolution representation. The HRMED_S model is based on the Channel-wise Knowledge Distillation approach to simplify the structure, achieve faster convergence, and refine the segmentation results by using conditional random fields instead of fully connected structures. The experimental results show that our system has better performance than other methods. The Intersection over the Union (IoU) of HRMED_T model reaches 0.756. The IoU of HRMED_S model also reaches 0.710 and params is only 3.99 M.

Preface

We are glad to welcome you to the proceedings of 2023 The 4th Asia Conference on Computers and Communications (ACCC 2023), which was held in Shanghai, China (virtual) during December 15-17, 2023. By following most of delegates’ willingness, the conference was entirely transferred as online conference instead of Shanghai, China as scheduled.ACCC 2023 welcomes contributions that involve computers and communications. All papers in this proceeding were subject to peer-review by conference committee members and international reviewers. Conference’s relevance and quality drive the selection of the papers included in the proceedings. The contributed papers presented in this volume cover a wide range of topics spanning sensor based electronic system design and circuit analysis, signal analysis and data transmission in future communication systems, modern intelligent system monitoring and security, and so on.The conference has lasted virtually for 3 days including the online testing, 2 keynote speeches, each with 40 minutes, 2 invited speeches, each with 30 minutes, as well as 3 oral sessions from December 15-17, 2023. All research presentations were split into 3 sessions with the theme below: Intelligent Detection and Image Analysis Based on Machine Vision, Next Generation Communication Systems and Information Security, Model based intelligent control system and simulation analysis. Each presenter has made 15-minute (including Q&A) oral presentation. The diverse interesting discussions give a generation of new ideas during conference. We are grateful that they share new research ideas with us.We are deeply grateful to committee members, and international reviewers, whose diligence and effort brought these proceedings successfully together. Their excellent professional advice enabled us to prepare this conference. Our special thanks go to our guest speakers for sharing their latest research results. We do hope these talks interest and inspire all of you.As you peruse these proceedings, we encourage you to explore the wealth of knowledge presented within. It is our hope that these research papers inspire further exploration and contribute to the ongoing development of computers and communications technologies. We look forward to future editions of ACCC, where the global community will continue to converge, share ideas, and shape the future of computing and communications.Thank you to all participants for making ACCC 2023 a memorable and intellectually stimulating event.ACCC 2023Organizing CommitteeList of Committees, Statement of Peer Review are available in this pdf.

Clinical evaluation of Artificial Intelligence Driven Remote Monitoring technology for assessment of patient oral hygiene during orthodontic treatment

This study aimed to clinically evaluate the accuracy of Dental Monitoring's (DM) artificial intelligence (AI) image analysis and oral hygiene notification algorithm in identifying oral hygiene and mucogingival conditions. Twenty-four patients seeking orthodontic therapy were monitored by DM oral hygiene protocol during their orthodontic treatment. During the bonding appointment and at each of 10 subsequent adjustment visits, a total of 232 clinical oral examinations were performed to assess the presence of the 3 oral hygiene parameters that DM monitors. In each clinical timepoint, the subjects took an oral DM scan and received a notification regarding their current oral status at that moment in time. Sensitivity, specificity, positive predictive value, andnegative predictive value were calculated to evaluate AI and clinical assessment of plaque, gingivitis, and recession. A total of 232 clinical time points have been evaluated clinically and by the DM AI algorithm. For DM's AI detection of plaque and calculus, gingivitis, and recession, the sensitivity was 0.53, 0.35, and 0.22; the specificity was 0.94, 0.96, and 0.99; and the accuracy was 0.60, 0.49, and 0.72, respectively. DM's oral hygiene notification algorithm has low sensitivity, high specificity, and moderate accuracy. This indicates a tendency of DM to underreport the presence of plaque, gingivitis, and recession.

Integration of Artificial Intelligence into Metallography: Area-wide Analysis of Microstructural Components of a Jominy Sample

Abstract Analysing the microstructure is an essential part of quality control in many steel manufacturing and processing operations. In this work, a promising method for autonomous analysis of microstructures in low-alloy steels based on artificial intelligence image analysis is presented. This study focuses on the classification of different microstructure components in metallographic images of steel microstructures using a Deep Convolutional Neural Network (DCNN) model. Since the accuracy of the model strongly depends on the size of the data set, a data set consisting of two million optical microscopy images was created to ensure the presence of different microstructure components and their combinations for training the system. The Jominy test was performed to verify the accuracy and capability of the microstructure analysis software. The AI makes it possible to analyse large amounts of image data with high precision and at the same time with less effort than conventional methods of microstructure components analysis.

Spine surgeon versus AI algorithm full-length radiographic measurements: a validation study of complex adult spinal deformity patients.

Spinal measurements play an integral role in surgical planning for a variety of spine procedures. Full-length imaging eliminates distortions that can occur with stitched images. However, these images take radiologists significantly longer to read than conventional radiographs. Artificial intelligence (AI) image analysis software that can make such measurements quickly and reliably would be advantageous to surgeons, radiologists, and the entire health system. Institutional Review Board approval was obtained for this study. Preoperative full-length standing anterior-posterior and lateral radiographs of patients that were previously measured by fellowship-trained spine surgeons at our institution were obtained. The measurements included lumbar lordosis (LL), greatest coronal Cobb angle (GCC), pelvic incidence (PI), coronal balance (CB), and T1-pelvic angle (T1PA). Inter-rater intra-class correlation (ICC) values were calculated based on an overlapping sample of 10 patients measured by surgeons. Full-length standing radiographs of an additional 100 patients were provided for AI software training. The AI algorithm then measured the radiographs and ICC values were calculated. ICC values for inter-rater reliability between surgeons were excellent and calculated to 0.97 for LL (95% CI 0.88-0.99), 0.78 (0.33-0.94) for GCC, 0.86 (0.55-0.96) for PI, 0.99 for CB (0.93-0.99), and 0.95 for T1PA (0.82-0.99). The algorithm computed the five selected parameters with ICC values between 0.70 and 0.94, indicating excellent reliability. Exemplary for the comparison of AI and surgeons, the ICC for LL was 0.88 (95% CI 0.83-0.92) and 0.93 for CB (0.90-0.95). GCC, PI, and T1PA could be determined with ICC values of 0.81 (0.69-0.87), 0.70 (0.60-0.78), and 0.94 (0.91-0.96) respectively. The AI algorithm presented here demonstrates excellent reliability for most of the parameters and good reliability for PI, with ICC values corresponding to measurements conducted by experienced surgeons. In future, it may facilitate the analysis of large data sets and aid physicians in diagnostics, pre-operative planning, and post-operative quality control.

Redefining comorbidity understanding in rheumatoid arthritis through novel approaches using real-world data

Rheumatoid arthritis (RA) is a prevalent chronic disease that is associated with numerous comorbidities. Accurate assessment of these coexisting conditions, as reported by clinicians, is critical for an improved understanding of the impact of the disease and patient care. This perspective aims to evaluate the utility of real-world data (RWD) for enhancing the understanding of comorbidities in RA and to assess its potential in reshaping clinical management. RWD approaches, specifically the use of structured databases or data extracted from electronic health records, offer promising alternatives to overcome the limitations of traditional methodologies. Structured databases provide a systematic approach to data analysis, utilizing diagnosis codes to study large patient cohorts, revealing the prevalence of conditions, and demonstrating the potential for long-term disease trend analysis. Meanwhile, natural language processing (NLP) and artificial intelligence (AI) image analysis can bridge the gap between structured and unstructured data, by extracting meaningful information from unstructured fields such as free text or imaging. NLP has proven effective in the identification of RA patients and research outcomes, while AI image analysis has enabled the discovery of hidden findings in cardiovascular assessments, establishing a basis for the assessment of comorbidities in RA. However, while the benefits of using RWD are substantial, challenges remain. Ensuring comprehensive data capture, managing missing data, and improving data detection are key areas requiring attention. The involvement of clinicians and researchers in rheumatology is crucial in unlocking the potential of RWD studies, offering the promise of significant improvements in disease characterization and patient health outcomes.

Automatic data augmentation for medical image segmentation using Adaptive Sequence-length based Deep Reinforcement Learning

Although existing deep reinforcement learning-based approaches have achieved some success in image augmentation tasks, their effectiveness and adequacy for data augmentation in intelligent medical image analysis are still unsatisfactory. Therefore, we propose a novel Adaptive Sequence-length based Deep Reinforcement Learning (ASDRL) model for Automatic Data Augmentation (AutoAug) in intelligent medical image analysis. The improvements of ASDRL-AutoAug are two-fold: (i) To remedy the problem of some augmented images being invalid, we construct a more accurate reward function based on different variations of the augmentation trajectories. This reward function assesses the validity of each augmentation transformation more accurately by introducing different information about the validity of the augmented images. (ii) Then, to alleviate the problem of insufficient augmentation, we further propose a more intelligent automatic stopping mechanism (ASM). ASM feeds a stop signal to the agent automatically by judging the adequacy of image augmentation. This ensures that each transformation before stopping the augmentation can smoothly improve the model performance. Extensive experimental results on three medical image segmentation datasets show that (i) ASDRL-AutoAug greatly outperforms the state-of-the-art data augmentation methods in medical image segmentation tasks, (ii) the proposed improvements are both effective and essential for ASDRL-AutoAug to achieve superior performance, and the new reward evaluates the transformations more accurately than existing reward functions, and (iii) we also demonstrate that ASDRL-AutoAug is adaptive for different images in terms of sequence length, as well as generalizable across different segmentation models.

A novel open-source CADs platform for 3D CT pulmonary analysis

Computer-aided diagnosis (CAD) systems play vital roles in the early detection of pulmonary nodules for reducing lung cancer mortality rates. To provide better services for professional doctors, this paper proposes an efficient open-source CAD platform with flexible equipments, user-friendly interfaces, and completed functions for 3D CT pulmonary nodule analysis. For the platform’s design and implementation, we fully consider application scenarios and system requirements. The platform supplies core functions for (1) Basic Image Processing, (2) Intelligent Image Analysis, (3) Multi-View Image Visualization, (4) Report Editing and Generation, (5) User Information Management, and (6) Inference Service Monitoring. Specifically, other state-of-the-art or user-defined algorithms can be integrated as plugin modules with no interference for system architecture. System evaluation with use-case testing demonstrates the effectiveness and universality of the proposed platform.

A comprehensive strategy of diabetic retinopathy screening in a public health system: Identifying and overcoming obstacles for implementation

AimsThis study reports the implementation of a diabetic retinopathy (DR) screening program in Aracaju, Brazil, emphasizing the challenges and premature termination. MethodsThe program, a collaboration between local health authorities, national DR screening experts, private clinics, trained retinal imaging technicians, a portable retinal camera with artificial intelligence (AI), telemedicine, and AI-assisted image analysis, was established. Screening occurred at primary care centers, with free specialized treatment for high-risk DR cases. ResultsAfter a public tender, two clinics were selected for screening, and four for secondary management. Initially, diabetic patient attendance at primary care clinics varied considerably, posing a challenge. Nonetheless, the screening continued, with 3,561 patients screened within six months. However, due to complaints and disputes between providers and authorities, the program was prematurely terminated, achieving less than a quarter of the initial screening target. ConclusionsThe DR screening program faced significant challenges, including low adherence and organizational issues at primary care centers. This experience highlights the need to standardize clinical workflows, enhance organization and integration among primary care units, and address cultural factors to improve adherence rates. These insights are valuable for implementing screening programs, particularly in low- and middle-income countries.

The use of rural cultural tourism product development strategy based on artificial intelligence technology

Abstract Based on artificial intelligence technology, this paper emphasizes user-centered tourism product development, focusing on the importance of meeting tourists’ core needs. The AI product development process is explained in detail, including the essential steps of intelligent image analysis and personalized recommendation. Tourist attractions are presented using image preprocessing algorithms, and intelligent image analysis is employed to analyze tourists’ interest preferences and create personalized image recommendations. Next, edge-aware intelligent visual guidance is employed to highlight cultural elements and target product development. Finally, the feasibility of the strategy for developing rural tourism products is confirmed. From January to December, the sales of tourism product development based on AI technology increased from 15,000 yuan to 38,200 yuan, which provides strong guidance for tourism operators.

Deep‐learning based in situ image monitoring crystal polymorph and size distribution: Modeling and validation

AbstractIn situ monitoring and closed‐loop control of the solution crystallization process are the modern trends for pharmaceutical development, in which the critical process parameters (CPPs) as well as the product critical quality attributes (CQAs) can be regulated and guaranteed during the manufacturing process. In this study, an in situ image monitoring methodology based on a state‐of‐the‐art deep‐learning model was developed to track the CQAs such as polymorph ratio, two‐dimensional crystal size, and crystal shape in a solvent‐mediated polymorphic transformation (SMPT) process. Coupled with the multidimensional process information, a 2D population balance model (PBM) was developed and validated using the results of the in situ image‐based CQAs analysis. The 2D‐PBM was solved using a high‐resolution finite volume method (HR‐FVM) which could provide a high dimensional particle‐size distribution. Through the validation between the process image analysis and the 2D‐PBM, the accuracy of image analysis was discussed, and the potential and challenges of in situ image analysis were proposed. This work aims to integrate the crystal polymorphism and two‐dimensional crystal size distribution (2D‐CSD) information in the SMPT process using intelligent microscopic image analysis and then to validate the results of neural network processing by solving the numerical solution of the multidimensional PBM.

EP07.02-04 The Prediction of Pathological Pleural Invasion of Lung Cancer by Artificial Intelligence Image Analysis

EP07.02-04 The Prediction of Pathological Pleural Invasion of Lung Cancer by Artificial Intelligence Image Analysis

Retracted: Intelligent Garment Graphic Design System for Artificial Intelligence and 3D Image Analysis

Retracted: Intelligent Garment Graphic Design System for Artificial Intelligence and 3D Image Analysis