Medical Imaging Research Articles

Colon Cancer Detection of Colonoscopy Images Using CNN

Abstract: Colon cancer requires an early and precise diagnosis in order to be effectively treated because it is a common and potentially fatal condition. This study proposes a reliable framework for colon cancer detection through image preprocessing and transfer learning approaches. The research compares the performance of three popular convolutional neural networks (CNNs) for this task: VGG-15, ResNet-50, and AlexNet.The image preprocessing phase entails a number of critical operations, including image conversion, resizing, and contrast enhancement using histogram equalization. In order to guarantee that the neural networks receive high-quality inputs, these preprocessing techniques strive to improve the quality and consistency of the colon cancer dataset. Pre-trained models extract and classify features from big datasets using transfer learning. The preprocessed colon cancer dataset is used to fine-tune the chosen CNN architectures VGG-15, ResNet-50, and AlexNet and adapt them for this particular medical imaging purpose.The results of this study show that transfer learning has the potential for colon cancer detection and provides helpful direction for researchers and practitioners looking for the best deep-learning architectures for early cancer diagnosis. This study adds to continuing efforts to improve the precision and effectiveness of colon cancer detection techniques, ultimately leading to better patient outcomes.

An Effective Classification of Brain Tumor using Deep Learning Techniques

Abstract: Brain tumor classification is a critical task in medical imaging, aiding in timely diagnosis and treatment planning. In this paper, we propose a comprehensive approach utilizing deep learning models for the classification of brain tumors into four categories: glioma, meningioma, pituitary tumors, and no tumors. We employ state-of-the-art convolutional neural network (CNN) architectures including ResNet50, DenseNet201, EfficientNetB3, and InceptionV3 for the classification task. To enhance the performance of the model, we employ resizing and augmentation techniques such as flips and rotation, thereby increasing the diversity of the training dataset. This is particularly crucial due to the limitations posed by small-sized datasets in previous methodologies. Our findings underscore the efficacy of deep learning approaches in brain tumor classification, with EfficientNetB3 emerging as a promising model for accurate diagnosis. Furthermore, our utilization of resizing and augmentation techniques demonstrates their significance in mitigating the challenges associated with limited training data.

Object Detection Model: An Enhanced SSD

Abstract: Object Detection plays a significant role in today’s era. Object detection plays a crucial role in various applications such as surveillance, medical images, and autonomous vehicles. The primary goal of object detection is to accurately detect the boundaries of objects of interest and classify those objects into predefined categories. In this research paper, we enhanced an object detection technique i.e. Single shot multibox detector (SSD) which is one of the top object detection technique in both aspect accuracy and speed. SSD is an object detection model that predicts object bounding boxes and class probabilities in a single forward pass. It uses multiple feature maps at different scales to cater to various object sizes and aspect ratios.The main contribution of this research is to enhance the accuracy, recall time, precision time and mean average time (mAP). The performance of this object detection technique is improved as the number of feature maps is increases, improved backbone, and enhanced activation functions, additional layers for enhanced detection and loss function modifications. An accuracy of 0.75 indicates that the model correctly predicts the class label for approximately 75% of all objects in the dataset. The enhanced model showed these results: a precision score of 0.75 implies that 75% of the objects identified by the model as positive (detected objects) are indeed true positives. A recall of 0.73 indicates that the model successfully identifies and detects 73% of all positive instances in the dataset. A mAP of 75.25 signifies the average precision across all classes, computed at different IOU thresholds (e.g., 0.5, 0.75).

Med-Tech-AI: Exploring Artificial Intelligence for Enhanced Healthcare Research

Abstract: This paper explores the potential of Artificial Intelligence (AI) to revolutionize healthcare research through the MedTech-AI project. The project investigates the development and evaluation of AI models for medical image analysis, aiming for improved disease detection and classification. It examines the potential of AI-powered clinical decision support systems (CDSS) to assist healthcare professionals. Additionally, the project explores Natural Language Processing (NLP) techniques for extracting valuable insights from unstructured healthcare data like electronic health records and medical literature. Finally, the paper investigates the use of AI for predictive analytics in disease prevention, identifying risk factors and informing preventative measures. By examining these areas, the Med-Tech-AI project establishes a foundation for understanding how AI can enhance healthcare research, potentially leading to improved healthcare efficiency, accuracy, and ultimately, better patient outcomes

Deep Learning Approach for Early Detection of Alzheimers Disease

Abstract: Alzheimer’s disease (AD) is a chronic, irreversible brain disorder, no effective cure for it till now. However, available medicines can delay its progress. Therefore, the early detection of AD plays a crucial role in preventing and controlling its progression. The main objective is to design an end-to-end framework for early detection of Alzheimer’s disease and medical image classification for various AD stages. A deep learning approach, specifically convolutional neural networks (CNN), is used in this work. Four stages of the AD spectrum are multi-classified. Furthermore, separate binary medical image classifications are implemented between each two-pair class of AD stages.

The Changing Landscape of Healthcare with State of the Art AI Technology

Abstract: Artificial Intelligence (AI) has emerged as a transformative force in healthcare, offering unparalleled opportunities to enhance patient outcomes, streamline processes, and reduce costs. This abstract provides an overview of the diverse applications of AI in healthcare, including medical imaging analysis, predictive analytics, personalized treatment plans, and administrative tasks automation. By leveraging machine learning algorithms and big data analytics, AI enables healthcare professionals to make more accurate diagnoses, predict disease progression, and customize treatment regimens tailored to individual patient needs. Furthermore, AI-powered tools facilitate the automation of routine administrative tasks, allowing healthcare providers to focus more on patient care. Despite the promising advancements, challenges such as data privacy concerns, regulatory hurdles, and the need for interdisciplinary collaboration remain. This abstract underscores the immense potential of AI in revolutionizing healthcare delivery while highlighting the importance of addressing associated ethical and regulatory considerations for its widespread adoption.

Comparative Analysis of VGG16, RESNET50, AND CNN Models for Lung Disease Prediction: A Deep Learning Approach

Abstract: Utilizing deep learning methods in medical image analysis has shown promise in enhancing disease detection and diagnosis. In this study, we conducted a detailed comparative analysis of threewidely recognized convolutional neural network (CNN) architectures: VGG16, ResNet50, and a custom CNN model tailored for pneumonia and COVID-19 detection from chest X-ray images. Leveraging transfer learning techniques and meticulously curated datasets, we evaluated the models' performance in accurately identifying respiratory diseases. Our investigation utilized two publicly available datasets, the ChestX-ray14 dataset and the NIH Chest X-ray Dataset, both annotated for pneumonia and COVID-19. Prior to model training, we conducted thorough preprocessing to ensure optimal data quality and consistency. Through rigorous experimentation, we assessed the models' accuracy, sensitivity, and specificity in disease detection. The results revealed nuanced differences in model performance across disease categories. While VGG16 demonstrated robust accuracy in pneumonia detection, ResNet50 exhibited enhanced sensitivity and specificity in identifying COVID-19 cases. Our custom CNN model, leveraging insights from both architectures, showcased competitive performance, emphasizing the importance of tailored model design for optimal diagnostic outcomes. Through comprehensive analysis and discussion, we elucidated the strengths and limitations of each model, considering factors such as computational efficiency, interpretability, and generalizability. Our findings underscore the potential of deep learningbased diagnostic tools in supporting healthcare professionals in timely and accurate disease diagnosis

Prediction of Abnormal Growth of Breast Cancer Cell Using Support Vector Machine

Abstract: Breast cancer is a serious health concern that affects millions of women worldwide. Early detection of breast cancer is crucial for effective treatment and improved survival rates. Artificial intelligence (AI) has shown great promise in the field of medical imaging and has been increasingly used for breast cancer detection. Cancer is one of the most dangerous diseases to humans, and yet no permanent cure has been developed for it. Breast cancer is one of the most common cancer types. According to the National Breast Cancer foundation, in 2020 alone, more than 276,000 new cases of invasive breast cancer and more than 48,000 non-invasive cases were diagnosed in the US. To put these figures in perspective, 64% of these cases are diagnosed early in the disease’s cycle, giving patients a 99% chance of survival. In this paper we develop a system for diagnosis, prognosis and prediction of breast cancer using AI. This will assist the doctors in diagnosis of the disease. In this research work, we systematically reviewed previous work done on detection and treatment of breast cancer using genetic sequencing or histopathological imaging with the help of deep learning and machine learning We use of AI for breast cancer detection, including studies that have used mammography, ultrasound, and magnetic resonance imaging (MRI) as imaging modalities. AI can also improve the accuracy and efficiency of breast cancer screening and diagnosis, as well as the challenges that must be addressed to ensure the successful integration of AI into clinical practice. However, further research and development are needed to optimize AI algorithms and ensure their safety and effectiveness in clinical practice. Breast cancer detection has emerged as a critical area of research, leveraging the power of Artificial Intelligence (AI) to enhance diagnostic accuracy and efficiency. Utilizing AI technologies, particularly deep learning, enables the development of robust and precise models for early breast cancer detection. These models analyze mammographic images, identifying subtle patterns indicative of malignancies that may escape human eye detection. Transfer learning, employing pre-trained neural network architectures, has demonstrated significant success in leveraging existing knowledge for improved performance. The integration of convolutional neural networks (CNNs) and image processing techniques empowers AI models to discern subtle abnormalities, aiding in the timely identification of potential breast tumors. Additionally, AI-driven diagnostic tools can streamline radiologists' workflow, providing rapid and reliable assessments, ultimately contributing to early intervention and improved patient outcomes. As these technologies continue to evolve, the synergy between AI and breast cancer detection promises to revolutionize screening methods, fostering a paradigm shift towards more accurate, efficient, and accessible diagnostic practices in the realm of breast health.

Computed Tomography- Based Maxillary Sinus Morphometry: A Tool for Gender Differentiation in Libyan Population

Introduction: In forensic contexts, gender identification is a multifaceted process that involves examining various anatomical characteristics to determine an individual's gender. Recent advancements in medical imaging techniques, such as computed tomography (CT) scans have provided researchers and healthcare professionals with more detailed insights into anatomical differences between genders such as the dimensions of paranasal sinuses. Aim: This study aimed to compare the maxillary sinus (MS) dimensions between genders in a Libyan sample using CT scans. Materials and methods: One hundred consecutive Libyan patients, comprising 50 women and 50 men, were included in the study. The Width, height, and length were determined based on CT images. Additionally, the perimeter, area, and volume of the MS were calculated mathematically. The collected data analyzed to compare the MS parameters between genders in both sides. Statistical significance was accepted at p<0.05. Results: Statistically we found that the length, width and length of the MS were significantly different between male & female in both sides, while there was no significant difference in case of sinus area and volume. The accuracy level for correctly classifying gender based on width of right MS was 91.8%, demonstrating strong predictive power of this parameter. The result showed also that 94 % of males and 90% of females were sexed correctly with an overall accuracy of 92%. Conclusion: This study highlights the importance of understanding sinus morphology in forensic anthropology for gender determination.

Artificial Intelligence and Machine Learning in Hepatocellular Carcinoma Screening, Diagnosis and Treatment - A Comprehensive Systematic Review

Background: Medical image analysis plays a crucial role in the screening, monitoring, diagnosis, and prognosis of diseases. Hepatocellular Carcinoma (HCC) often remain asymptomatic in their early stages. Timely identification and intervention are crucial to prevent the progression to decompensated liver diseases and advanced-stage HCC, minimizing morbidity and mortality. Methodology: This study examined twenty relevant research articles that met the inclusion criteria. Utilizing the PRISMA criteria, a systematic search was conducted on PubMed/MEDLINE and Google Scholar for studies on HCC screening employing Artificial Intelligence (AI) and Machine Learning (ML). The search focused on the keywords "artificial intelligence" and "hepatocellular carcinoma," with inclusion criteria specifying studies in the English language published in and after 2020. Exclusions were made for histology, animal research, and investigations conducted before 2020. Titles and abstracts were thoroughly reviewed, and any discrepancies were discussed. Results: The comprehensive review reveals the transformative impact of AI and ML on HCC screening, diagnosis, and therapy. ML models demonstrated effectiveness in early HCC diagnosis, distinguishing hepatic lesions, predicting treatment responses, and assessing recurrence risks across various techniques. Integration of mass spectrometry-based technologies, advanced imaging, and real-world data significantly improved diagnostic accuracy and clinical decision support. AI models exhibited diagnostic expertise with potential applications in therapy suggestions, personalized surveillance, and prognostic evaluations. However, further validation and seamless integration into clinical practice are essential for realizing their full potential. Conclusion: This systematic study underscores the progress made in the application of AI and ML in HCC screening, diagnosis, and treatment. Numerous studies highlight the capability of AI and ML systems to enhance hepatocellular carcinoma diagnosis, predict outcomes, and optimize therapy. The enduring and versatile nature of these technologies points towards a revolutionary future in personalized and efficient HCC management.

High-Resolution X-ray Image from Copper-Based Perovskite Hybrid Polymer.

Cs3Cu2I5 nanocrystals (NCs) are considered to be promising materials due to their high photoluminescence efficiency, lack of lead toxicity, and X-ray responsiveness. However, during the crystallization process, NCs are prone to agglomeration and exhibit uneven size distribution, resulting in several light scattering that severely affect their imaging resolution. Herein, we successfully developed a high-resolution scintillator film by growing copper-based perovskite NCs within a hybrid polymer matrix. By leveraging the ingenious integration of polyvinylidene fluoride (PVDF) and polymethyl methacrylate (PMMA), the size and distribution uniformity of Cs3Cu2I5 NCs can be effectively controlled. Consequently, a high spatial resolution of 14.3 lp mm-1 and a low detection limit of 105 nGy s-1 are achieved, and the scintillator film has excellent flexibility and stability. These results highlight the promising application of Cs3Cu2I5 scintillator films in low-cost, flexible, and high-performance medical imaging.

Hi-gMISnet: generalized medical image segmentation using DWT based multilayer fusion and dual mode attention into high resolution pGAN

Objective. Automatic medical image segmentation is crucial for accurately isolating target tissue areas in the image from background tissues, facilitating precise diagnoses and procedures. While the proliferation of publicly available clinical datasets led to the development of deep learning-based medical image segmentation methods, a generalized, accurate, robust, and reliable approach across diverse imaging modalities remains elusive. Approach. This paper proposes a novel high-resolution parallel generative adversarial network (pGAN)-based generalized deep learning method for automatic segmentation of medical images from diverse imaging modalities. The proposed method showcases better performance and generalizability by incorporating novel components such as partial hybrid transfer learning, discrete wavelet transform (DWT)-based multilayer and multiresolution feature fusion in the encoder, and a dual mode attention gate in the decoder of the multi-resolution U-Net-based GAN. With multi-objective adversarial training loss functions including a unique reciprocal loss for enforcing cooperative learning in pGANs, it further enhances the robustness and accuracy of the segmentation map. Main results. Experimental evaluations conducted on nine diverse publicly available medical image segmentation datasets, including PhysioNet ICH, BUSI, CVC-ClinicDB, MoNuSeg, GLAS, ISIC-2018, DRIVE, Montgomery, and PROMISE12, demonstrate the proposed method’s superior performance. The proposed method achieves mean F1 scores of 79.53%, 88.68%, 82.50%, 93.25%, 90.40%, 94.19%, 81.65%, 98.48%, and 90.79%, respectively, on the above datasets, surpass state-of-the-art segmentation methods. Furthermore, our proposed method demonstrates robust multi-domain segmentation capabilities, exhibiting consistent and reliable performance. The assessment of the model’s proficiency in accurately identifying small details indicates that the high-resolution generalized medical image segmentation network (Hi-gMISnet) is more precise in segmenting even when the target area is very small. Significance. The proposed method provides robust and reliable segmentation performance on medical images, and thus it has the potential to be used in a clinical setting for the diagnosis of patients.

Boosting wisdom of the crowd for medical image annotation using training performance and task features.

A crucial bottleneck in medical artificial intelligence (AI) is high-quality labeled medical datasets. In this paper, we test a large variety of wisdom of the crowd algorithms to label medical images that were initially classified by individuals recruited through an app-based platform. Individuals classified skin lesions from the International Skin Lesion Challenge 2018 into 7 different categories. There was a large dispersion in the geographical location, experience, training, and performance of the recruited individuals. We tested several wisdom of the crowd algorithms of varying complexity from a simple unweighted average to more complex Bayesian models that account for individual patterns of errors. Using a switchboard analysis, we observe that the best-performing algorithms rely on selecting top performers, weighting decisions by training accuracy, and take into account the task environment. These algorithms far exceed expert performance. We conclude by discussing the implications of these approaches for the development of medical AI.

Application of CNN Model on Medical Images (Chest X-Rays)

Chest diseases encompass a range of conditions affecting the lungs, including infections caused by bacteria, viruses, or other microorganisms. These diseases can vary in severity, posing a potential threat to vulnerable individuals such as young children, older adults, and those with weakened immune systems. Detecting chest diseases involves employing diverse methods, such as physical examination, chest x-rays, blood tests, and sputum culture. During a physical examination, healthcare professionals utilize a stethoscope to listen for abnormal breathing sounds or crackling noises in the chest. They also assess factors like fever and respiratory rate. Chest x-rays provide a visual representation of the lungs, enabling the identification of areas displaying inflammation or fluid accumulation. Sputum culture involves collecting a sample of mucus expelled from the lungs through coughing and testing it for the presence of microorganisms. Key Words: Deep Learning, CNN, Chest X-Ray Images.

Duodenal-Jejunal Stenosis as a Manifestation of Eosinophilic Gastroenteritis: A Case Report

Introduction: Eosinophilic gastroenteritis (EGE) is a rare disorder characterized by eosinophilic infiltration of the stomach and small intestine. Its clinical manifestations depend on the extent, location, and depth of the inflammatory infiltrate. The pathogenesis remains unclear, but it may be related to food allergy or other hypersensitivity reactions in some patients. Untreated cases can cause intestinal stricture or perforation, making timely diagnosis and intervention essential. Clinical case: A 13-year-old male adolescent was hospitalized due to symptoms suggesting intestinal obstruction or alterations in intestinal motility. Medical imaging revealed significant gastric distension and delayed gastric emptying. An endoscopic study showed a “snake-skin” pattern, duodenitis due to food stasis, a 2 cm ulcer, and stenosis in the duodenum. The biopsy confirmed EGE with significant irregular eosinophilic infiltration of the lamina propria. A panendoscopy was performed after steroid treatment without being able to assess the site of stenosis and perform pneumatic dilation. After completing the immunomodulatory management, a contrast-enhanced abdominal tomography was performed showed a significant decrease in the intestinal lumen, therefore an exploratory laparotomy with a lateral to lateral jejune-jejunal anastomosis was performed. Discussion: EGE is a rare inflammatory disorder characterized by eosinophilic infiltration of the stomach and small intestine. Symptoms are based on the affected gastrointestinal tract segment. In this case, our patient presented intestinal obstruction symptoms with endoscopic biopsies and a histopathological study confirmed eosinophil infiltration. Treatment typically involves identifying potential triggers and corticosteroids to reduce inflammation and symptoms. However, in this case, a decrease in the intestinal lumen was observed, therefore an elective surgical resection was performed. Conclusion: Eosinophilic gastroenteritis is a rare disorder that requires strong clinical suspicion to be diagnosed, which is why its knowledge is important. In this case, gastrointestinal symptoms, radiological and endoscopic images, in addition to the histopathological report, were key tools for integrating the diagnosis. Although management with steroids and immunomodulators was not successful, surgical management became an appropriate option as definitive treatment.

Deep Learning Analysis of 3D Volume of Brain Anomaly Using MRI

In this study, we present a comprehensive investigation into the efficacy of two state-of-the-art deep learning architectures, 3D UNet and VNet, for brain tumor detection and segmentation in volumetric medical images.

Production of carbon-11 for PET preclinical imaging using a high-repetition rate laser-driven proton source

Most advanced medical imaging techniques, such as positron-emission tomography (PET), require tracers that are produced in conventional particle accelerators. This paper focuses on the evaluation of a potential alternative technology based on laser-driven ion acceleration for the production of radioisotopes for PET imaging. We report for the first time the use of a high-repetition rate, ultra-intense laser system for the production of carbon-11 in multi-shot operation. Proton bunches with energies up to 10–14 MeV were systematically accelerated in long series at pulse rates between 0.1 and 1 Hz using a PW-class laser. These protons were used to activate a boron target via the 11\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$^{11}$$\\end{document}B(p,n)11\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$^{11}$$\\end{document}C nuclear reaction. A peak activity of 234 kBq was obtained in multi-shot operation with laser pulses with an energy of 25 J. Significant carbon-11 production was also achieved for lower pulse energies. The experimental carbon-11 activities measured in this work are comparable to the levels required for preclinical PET, which would be feasible by operating at the repetition rate of current state-of-the-art technology (10 Hz). The scalability of next-generation laser-driven accelerators in terms of this parameter for sustained operation over time could increase these overall levels into the clinical PET range.

Early detection of glaucoma: feature visualization with a deep convolutional network

ABSTRACT Computer-aided diagnosis of ocular disorders like glaucoma can be effectively performed with retinal fundus images. Since the advent of machine learning and later deep learning techniques in medical image analysis, research focusing on automated and early detection of glaucoma has gained prominence. In this paper, we show how a deep Convolutional Neural Network model can be used in visualising the features related to Glaucoma that are consistent with (a subset of) those used by ophthalmologists for diagnosis. However, since the doctors do not necessarily depend entirely on these features but back them up with other important measurements such as visual field testing (HVF), any deep neural network such as CNN may not necessarily be expected to yield 100% accuracy. Further, other important parameters such as Intraocular Pressure (IOP), which may be built into a deep learning model, may not necessarily correlate well with the presence of Glaucoma. In this work, a deep CNN model has been trained and tested on a large number of high-quality fundus images, both normal and glaucomatous. We compare the results with transfer learning models such as VGG16, ResNet50, and MobileNetV2. On an average, we obtained an accuracy of 93.75% in identifying glaucoma, focusing only on the features of the fundus image.

Enhancing EfficientNetv2 with global and efficient channel attention mechanisms for accurate MRI-Based brain tumor classification

AbstractThe early and accurate diagnosis of brain tumors is critical for effective treatment planning, with Magnetic Resonance Imaging (MRI) serving as a key tool in the non-invasive examination of such conditions. Despite the advancements in Computer-Aided Diagnosis (CADx) systems powered by deep learning, the challenge of accurately classifying brain tumors from MRI scans persists due to the high variability of tumor appearances and the subtlety of early-stage manifestations. This work introduces a novel adaptation of the EfficientNetv2 architecture, enhanced with Global Attention Mechanism (GAM) and Efficient Channel Attention (ECA), aimed at overcoming these hurdles. This enhancement not only amplifies the model’s ability to focus on salient features within complex MRI images but also significantly improves the classification accuracy of brain tumors. Our approach distinguishes itself by meticulously integrating attention mechanisms that systematically enhance feature extraction, thereby achieving superior performance in detecting a broad spectrum of brain tumors. Demonstrated through extensive experiments on a large public dataset, our model achieves an exceptional high-test accuracy of 99.76%, setting a new benchmark in MRI-based brain tumor classification. Moreover, the incorporation of Grad-CAM visualization techniques sheds light on the model’s decision-making process, offering transparent and interpretable insights that are invaluable for clinical assessment. By addressing the limitations inherent in previous models, this study not only advances the field of medical imaging analysis but also highlights the pivotal role of attention mechanisms in enhancing the interpretability and accuracy of deep learning models for brain tumor diagnosis. This research sets the stage for advanced CADx systems, enhancing patient care and treatment outcomes.

Analytical Study of Medical PET Images by DWT and PCA

The PET medical images need clarification to reach a correct diagnosis of the disease by removing noise from them, and this is done by applying several techniques, including DWT, PCA. DWT is an algorithm that compresses an image and removes noise by dividing the image into several levels, removing the unnecessary (neglected) part of the image. The PCA algorithm compresses the error parameters in the image and then reconstructs them.