Computer-aided Diagnosis Research Articles

Automated Bone Tumor Detection Using Deep Learning: A CNN-Based Approach For Enhanced Diagnostic Accuracy

Abstract - Primary bone tumors present significant diagnostic challenges on radiographs, often requiring specialized expertise for accurate and timely identification.1 Early detection is crucial for a favorable prognosis, particularly for malignant types, which represent a leading cause of cancer-related mortality in adolescents and young adults.3 This study develops and evaluates a deep learning (DL) model, specifically Faster R-CNN with a ResNet backbone, for the automated detection and classification (benign vs. malignant) of primary bone tumors on radiographs. The model was trained and validated using the publicly available Bone Tumor X-ray Radiograph (BTXRD) dataset. The DL model demonstrates significant potential as an assistive tool for radiologists in detecting and classifying primary bone tumors on radiographs, potentially improving diagnostic accuracy and efficiency, particularly in non-specialized settings. Key Words: Bone Tumor, Deep Learning, Radiography, X-ray, Object Detection, Classification, Convolutional Neural Network (CNN), Faster R-CNN, BTXRD, Computer-Aided Diagnosis.

AUTOMATED LUNG CANCER DETECTION USING NAS: A HIGH-PERFORMANCE DEEP LEARNING APPROACH

Abstract – Lung cancer remains one of the leading causes of mortality worldwide, necessitating early and accurate detection for effective treatment. This study explores the application of deep learning techniques, specifically Convolutional Neural Networks (CNNs) and Neural Architecture Search (NAS), for automated lung cancer detection from CT scan images. CNNs, while effective, often require manual architecture tuning, leading to suboptimal performance. NAS, on the other hand, optimizes network architecture automatically, resulting in improved accuracy. Experimental results demonstrate that CNN achieves an accuracy of 84.38%, whereas NAS significantly outperforms it with an accuracy of 96.35%. The superior performance of NAS is attributed to its ability to discover the most efficient network structure tailored to lung cancer detection. These findings highlight the potential of automated deep learning approaches in medical image analysis, contributing to more reliable and precise diagnostic tools. Keywords: Lung Cancer Detection, Deep Learning, Neural Architecture Search (NAS), Convolutional Neural Networks (CNN).

Designing a web-based application for computer-aided diagnosis of intraosseous jaw lesions and assessment of its diagnostic accuracy.

This study aimed to design a web-based application for computer-aided diagnosis (CADx) of intraosseous jaw lesions, and assess its diagnostic accuracy. In this diagnostic test study, a web-based application was designed for CADx of 19 types of intraosseous jaw lesions. To assess its diagnostic accuracy, clinical and radiographic information of 95 cases with confirmed histopathological diagnosis of intraosseous jaw lesions was retrieved from hospital archives and published literature and imported to the application by a senior dental student. The top-N accuracy, kappa value, and Brier score were calculated, and the sensitivity, specificity, positive (PPV) and negative (NPV) predictive values, and the area under the receiver operating characteristic (ROC) curve (AUC) were calculated separately for each lesion according to DeLong et al. were calculated separately for each lesion according to DeLong et al. In assessment of top-N accuracy, the designed application gave a correct differential diagnosis in 93 cases (97.89%); the correct diagnosis was at the top of the list of differential diagnoses in 78 cases (82.10%); these values were 85 (89.47%) and 67 (70.52%) for an oral radiologist. The kappa value was 0.53. The Brayer score for the prevalence match was 0.18, and the pattern match was 0.15. The results highlighted the optimally high diagnostic accuracy of the designed application, indicating that it may be reliably used for CADx of intraosseous jaw lesions, if given accurate data.

DC-TransDPANet: A Transformer-Based Framework Integrating Composite Attention and Polarized Attention for Medical Image Segmentation

Medical image segmentation is a critical task in image analysis and plays an essential role in computer-aided diagnosis. Despite the promising performance of hybrid models combining U-Net and transformer architectures, these approaches face challenges in extracting local features and optimizing attention mechanisms. To address these limitations, we propose the Depthwise Composite Transformer and Depthwise Polarized Attention Network (DC-TransDPANet), a novel framework designed for medical image segmentation. The proposed DC-TransDPANet introduces a Depthwise Composite Attention Module (DW-CAM), which integrates depthwise convolution, and a Composite Attention mechanism to enhance local feature extraction and fuse contextual information. Additionally, a Depthwise Polarized Attention (DPA) block is employed to improve global context representation while preserving high-resolution details, achieving a fine balance between local and global feature extraction. Extensive experiments on benchmark datasets demonstrate that DC-TransDPANet significantly outperforms existing methods in segmentation accuracy.

A hybrid AI method for lung cancer classification using explainable AI techniques.

A hybrid AI method for lung cancer classification using explainable AI techniques.

Maize Leaf Disease Prediction Using Deep Learning

This research addresses the critical challenge of maize crop diseases by leveraging Convolutional Neural Networks (CNNs) for automated disease detection, overcoming the inefficiencies of traditional manual methods. The proposed CNN-based framework effectively classifies diseases such as Gray Leaf Spot, Common Rust, and Northern Leaf Blight by incorporating advanced preprocessing techniques like data augmentation and normalization to enhance accuracy and robustness. By enabling early disease detection, the system supports timely interventions, reduces crop losses, and promotes sustainable farming practices. Future directions include expanding datasets, integrating IoT for real-time monitoring, and exploring advanced DL architectures to further optimize performance, contributing to global food security and modernized agriculture.

Artificial intelligence applications for the diagnosis of pulmonary nodules.

This review evaluates the role of artificial intelligence (AI) in diagnosing solitary pulmonary nodules (SPNs), focusing on clinical applications and limitations in pulmonary medicine. It explores AI's utility in imaging and blood/tissue-based diagnostics, emphasizing practical challenges over technical details of deep learning methods. AI enhances computed tomography (CT)-based computer-aided diagnosis (CAD) through steps like nodule detection, false positive reduction, segmentation, and classification, leveraging convolutional neural networks and machine learning. Segmentation achieves Dice similarity coefficients of 0.70-0.92, while malignancy classification yields areas under the curve of 0.86-0.97. AI-driven blood tests, incorporating RNA sequencing and clinical data, report AUCs up to 0.907 for distinguishing benign from malignant nodules. However, most models lack prospective, multiinstitutional validation, risking overfitting and limited generalizability. The "black box" nature of AI, coupled with overlapping inputs (e.g., nodule size, smoking history) with physician assessments, complicates integration into clinical workflows and precludes standard Bayesian analysis. AI shows promise for SPN diagnosis but requires rigorous validation in diverse populations and better clinician training for effective use. Rather than replacing judgment, AI should serve as a second opinion, with its reported performance metrics understood as study-specific, not directly applicable at the bedside due to double-counting issues.

3D Reconstruction and Large-Scale Detection of Roads Based on UAV Imagery

Accurate and efficient detection of road damage is crucial in traffic safety and maintenance management. Traditional road detection methods have problems such as low efficiency and insufficient accuracy, making it difficult to meet the needs of large-scale road health assessments. With the development of drone technology and computer vision, new ideas have been provided for the automatic detection of road diseases. The existing drone-based road detection methods have poor performance in dealing with complex road scenes such as vehicle occlusion, and there is still room for improvement in 3D modeling accuracy and disease detection accuracy, lacking a comprehensive and efficient solution. This paper proposes a UAV (Unmanned Aerial Vehicle)-based 3D reconstruction and large-scale disease detection method for roads. By capturing aerial images with UAVs and utilizing an improved YOLOv8 model, vehicles in the images are identified and removed. Apply MVSNet (Multi-View Stereo Network) 3D reconstruction algorithm for road surface modeling, and finally use point cloud processing and DBSCAN (Density-Based Spatial Clustering of Applications with Noise) clustering for disease detection. The experimental results show that this method performs excellently in terms of 3D modeling accuracy and speed. Compared with the traditional colmap method, the reconstruction speed is greatly improved, and the reconstruction density is three times that of colmap. Meanwhile, the reconstructed point cloud can effectively detect road smoothness and settlement. This study provides a new method for effective disease detection under complex road conditions, suitable for large-scale road health assessment tasks.

Explainable Artificial Intelligence for Diagnosis and Staging of Liver Cirrhosis Using Stacked Ensemble and Multi-Task Learning

Background/Objectives: Liver cirrhosis is a critical chronic condition with increasing global mortality and morbidity rates, emphasizing the necessity for early and accurate diagnosis. This study proposes a comprehensive deep-learning framework for the automatic diagnosis and staging of liver cirrhosis using T2-weighted MRI images. Methods: The methodology integrates stacked ensemble learning, multi-task learning (MTL), and transfer learning within an explainable artificial intelligence (XAI) context to improve diagnostic accuracy, reliability, and transparency. A hybrid model combining multiple pre-trained convolutional neural networks (VGG16, MobileNet, and DenseNet121) with XGBoost as a meta-classifier demonstrated robust performance in binary classification between healthy and cirrhotic cases. Results: The model achieved a mean accuracy of 96.92%, precision of 95.12%, recall of 98.93%, and F1-score of 96.98% across 10-fold cross-validation. For staging (mild, moderate, and severe), the MTL framework reached a main task accuracy of 96.71% and an average AUC of 99.81%, with a powerful performance in identifying severe cases. Grad-CAM visualizations reveal class-specific activation regions, enhancing the transparency and trust in the model’s decision-making. The proposed system was validated using the CirrMRI600+ dataset with a 10-fold cross-validation strategy, achieving high accuracy (AUC: 99.7%) and consistent results across folds. Conclusions: This research not only advances State-of-the-Art diagnostic methods but also addresses the black-box nature of deep learning in clinical applications. The framework offers potential as a decision-support system for radiologists, contributing to early detection, effective staging, personalized treatment planning, and better-informed treatment planning for liver cirrhosis.

Domain knowledge-infused pre-trained deep learning models for efficient white blood cell classification

White blood cell (WBC) classification is a crucial step in assessing a patient’s health and validating medical treatment in the medical domain. Hence, efficient computer vision solutions to the classification of WBC will be an effective aid to medical practitioners. Computer-aided diagnosis (CAD) reduces manual intervention, avoids errors, speeds up medical analysis, and provides accurate medical reports. Though a lot of research has been taken up to develop deep learning models for efficient classification of WBCs, there is still scope for improvement to support the data insufficiency issue in medical data sets. Data augmentation and normalization techniques increase the quantity of data but don’t enhance the quality of the data. Hence, deep learning models though performing well can still be made efficient and effective when quality data is fused along with the available image dataset. This paper aims to utilize domain knowledge and image data to improve the classification performance of pre-trained models namely Inception V3, DenseNet 121, ResNet 50, MobileNet V2, and VGG 16. The models performance, with and without domain knowledge infused, is analyzed on the BCCD and LISC datasets. On the BCCD dataset, the average accuracies increased from 82.7%, 98.8%, 98.38%, 98.56%, and 98.5%–99.38%, 99.05%, 99.05%, 98.67%, and 98.75% for Inception V3, DenseNet 121, ResNet 50, MobileNet V2, and VGG 16, respectively. Similarly, on the LISC dataset, the accuracies improved from 86.76%, 92.2%, 91.76%, 92.8%, and 94.4%–92.05%, 95.88%, 95.58%, 95.2%, and 95.2%, respectively.

Plant Leaf Disease Detection System Using CNN

Abstract: Fruit and vegetable crops experience diminished agricultural production output because of pests along with diseases that rank as major factors worldwide. The correct identification of these issues becomes vital because delayed detection results in decreased quantity and quality of yields which then causes problems for food supply networks and regional economic stability. Farmers traditionally monitor plant diseases through individual observation assisted by expert consultations because they search for clear indicators of leaf damage including discolorations or spotted lesions or deteriors on leaves. The approach fails to meet standards because it often produces unsuitable results and unreliable human involvement. The proposed deep learning-based Disease Recognition Model employs Convolutional Neural Networks (CNNs) for processing leaf disease diagnosis within apple and corn and tomato and potato crops. The system enables automatic disease detection of leaves through image processing which delivers precise results. The training data consists of multiple leaf images which come from healthy subjects and disease-infected samples enabling precise identification of various diseases. The tool aims to become an affordable solution that supports farmers and agronomists and policymakers for better crop health management and minimal chemical usage while ensuring sustainable farming practices . Key Words: The system employs key terms including Leaf disease detection, plant health monitoring, CNN classification, fruit and vegetable crops, automated diagnosis, early disease intervention, sustainable agriculture, precision farming.

One-Dimensional Convolutional Neural Network for Automated Kimchi Cabbage Downy Mildew Detection Using Aerial Hyperspectral Images

Downy mildew poses a significant threat to kimchi cabbage, a vital agricultural product in Korea, adversely affecting its yield and quality. Traditional disease detection methods based on visual inspection are labor intensive and time consuming. This study proposes a non-destructive, field-scale disease detection approach using unmanned aerial vehicle (UAV)-based hyperspectral imaging. Hyperspectral images of the kimchi cabbage field were preprocessed, segmented at the pixel level, and classified into four categories: background, healthy, early-stage disease, and late-stage disease. Spectral analysis of the late and early stages of downy mildew infection revealed notable differences in the red-edge band, with infected plants exhibiting increased red-edge reflectance. To automate disease detection, various machine learning models, including Random Forest (RF), 1D Convolutional Neural Network (1D-CNN), 1D Residual Network (1D-ResNet), and 1D Inception Network (1D-InceptionNet), were developed. These models were trained based on a 0.2 sampling dataset, achieving overall accuracy scores of 0.907, 0.901, 0.909, and 0.914, along with F1 scores of 0.876, 0.845, 0.897, and 0.899, respectively. Overall, the results of this study revealed that the red-edge band reliably signaled the presence of downy mildew, and the 1D-InceptionNet model demonstrated the most effective performance for automatic disease detection.

International Journal of Scientific Research in Computer Science, Engineering and Information Technology

Traditional skin disease diagnosis is often slow, expensive, and requires in-person consultations, making it less accessible for many individuals. Conventional Computer-Aided Diagnosis (CAD) methods rely on manually extracted features like color, texture, and shape, which limits their accuracy, particularly across diverse skin tones. Additionally, online symptom checkers and existing AI models often lack real-time processing capabilities and mobile accessibility, reducing their effectiveness in providing instant and accurate results. To address these limitations, we developed Dr.Advice, an AI-powered Android application designed for real-time skin disease detection. Built with Python, Java, C++, Kotlin, and Android Studio SDK, Dr.Advice integrates advanced machine learning techniques to analyze skin images, detect conditions, and provide diagnostic insights. The application features real-time image processing, a user-friendly interface, and secure data handling, ensuring privacy and accuracy. Trained on diverse datasets, the model enhances detection accuracy across various skin tones. By offering fast, reliable, and accessible early diagnosis, Dr.Advice aims to revolutionize dermatological care, improving treatment outcomes and making skin disease detection more efficient and widely available.

Multimodal GPT model for assisting thyroid nodule diagnosis and management

Although using artificial intelligence (AI) to analyze ultrasound images is a promising approach to assessing thyroid nodule risks, traditional AI models lack transparency and interpretability. We developed a multimodal generative pre-trained transformer for thyroid nodules (ThyGPT), aiming to provide a transparent and interpretable AI copilot model for thyroid nodule risk assessment and management. Ultrasound data from 59,406 patients across nine hospitals were retrospectively collected to train and test the model. After training, ThyGPT was found to assist in reducing biopsy rates by more than 40% without increasing missed diagnoses. In addition, it detects errors in ultrasound reports 1,610 times faster than humans. With the assistance of ThyGPT, the area under the curve for radiologists in assessing thyroid nodule risks improved from 0.805 to 0.908 (p < 0.001). As an AI-generated content-enhanced computer-aided diagnosis (AIGC-CAD) model, ThyGPT has the potential to revolutionize how radiologists use such tools.

A Hybrid Deep Learning Model to Improve Skin Cancer Classification

Abstract—Among the various types of cancer globally, skin cancer is one of the most common, and identifying it early is vital for improving patient outcomes. This paper presents a novel hybrid deep learning approach for categorizing skin lesions into seven distinct groups using the HAM10000 dataset. We propose and compare two architectures: a CNN-DenseNet121 hybrid model and a CNN-ResNet50V2 hybrid model with an attention mechanism. Our experimental results demonstrate that the CNN-DenseNet121 model achieves superior overall accuracy, while the attention-based model shows improved focus on diagnostically relevant regions and better performance on certain challenging categories. The integration of attention mechanisms with transfer learning provides a more focused feature extraction process, which is essential for distinguishing subtle differences between benign and malignant skin conditions. This research contributes to the ongoing development of computer-aided diagnosis systems for dermatological applications. Index Terms—Neural networks, deep learning, attention models, dermatological cancer, classification of medical images, knowledge transfer

Unified Multi-Modal Diagnostic Framework With Reconstruction Pre-Training and Heterogeneity-Combat Tuning.

Medical multi-modal pre-training has revealed promise in computer-aided diagnosis by leveraging large-scale unlabeled datasets. However, existing methods based on masked autoencoders mainly rely on data-level reconstruction tasks, but lack high-level semantic information. Furthermore, two significant heterogeneity challenges hinder the transfer of pre-trained knowledge to downstream tasks, i.e., the distribution heterogeneity between pre-training data and downstream data, and the modality heterogeneity within downstream data. To address these challenges, we propose a Unified Medical Multi-modal Diagnostic (UMD) framework with tailored pre-training and downstream tuning strategies. Specifically, to enhance the representation abilities of vision and language encoders, we propose the Multi-level Reconstruction Pre-training (MR-Pretrain) strategy, including a feature-level and data-level reconstruction, which guides models to capture the semantic information from masked inputs of different modalities. Moreover, to tackle two kinds of heterogeneities during the downstream tuning, we present the heterogeneity-combat downstream tuning strategy, which consists of a Task-oriented Distribution Calibration (TD-Calib) and a Gradient-guided Modality Coordination (GM-Coord). In particular, TD-Calib fine-tunes the pre-trained model regarding the distribution of downstream datasets, and GM-Coord adjusts the gradient weights according to the dynamic optimization status of different modalities. Extensive experiments on five public medical datasets demonstrate the effectiveness of our UMD framework, which remarkably outperforms existing approaches on three kinds of downstream tasks.

FEGGNN: Feature-Enhanced Gated Graph Neural Network for robust few-shot skin disease classification.

FEGGNN: Feature-Enhanced Gated Graph Neural Network for robust few-shot skin disease classification.

Melanoma Skin Classification Using the Hybrid Approach Residual Network-Vision Transformer for Cancer Diagnosis.

Computer-aided diagnosis using deep neural networks allows for the analysis and processing of images or videos of different pathologies, providing valuable reference data to physicians for the diagnosis or screening of conditions such as skin cancer. In this work, we highlight the contribution of Convolutional Neural Networks, pre-trained models, and Vision Transformer architectures in the classification of skin melanoma. The experimental aspect will therefore involve the contribution of the classical CNN, as well as models inspired by this CNN, namely, Inception V3, ResNet 50, AlexNet, and EfficientNet in addition to the hybrid architecture. The conducted experiments entailed the adjustment of multiple hyperparameters, leading to the development of an architecture that achieved optimal results. Additionally, employing a hybrid architecture not only facilitated the amalgamation of the strengths from two models (the top performing pretrained ResNet50 model with the Vision Transformer) but also led to enhanced accuracy. After training the dataset, the proposed models have contributed to progressively improving the results, eventually achieving a classification rate of 95.53% for the hybrid ResNet50-ViT model. The aim of this research is to equip clinicians with a robust tool for melanoma diagnosis by leveraging the strengths of two models within the ResNet50-ViT hybrid framework.

Deep learning for cerebral vascular occlusion segmentation: A novel ConvNeXtV2 and GRN-integrated U-Net framework for diffusion-weighted imaging.

Deep learning for cerebral vascular occlusion segmentation: A novel ConvNeXtV2 and GRN-integrated U-Net framework for diffusion-weighted imaging.

Inspired by “Focus, Fusion, Collaboration”: A multi-level ensemble network for automatic pneumonia diagnosis from full slice CT images

Inspired by “Focus, Fusion, Collaboration”: A multi-level ensemble network for automatic pneumonia diagnosis from full slice CT images