Custom Convolutional Neural Network Research Articles

Computer-aided cholelithiasis diagnosis using explainable convolutional neural network.

Accurate and precise identification of cholelithiasis is essential for saving the lives of millions of people worldwide. Although several computer-aided cholelithiasis diagnosis approaches have been introduced in the literature, their use is limited because Convolutional Neural Network (CNN) models are black box in nature. Therefore, a novel approach for cholelithiasis classification using custom CNN with post-hoc model explanation is proposed. This paper presents multiple contributions. First, a custom CNN architecture is proposed to classify and predict cholelithiasis from ultrasound image. Second, a modified deep convolutional generative adversarial network is proposed to produce synthetic ultrasound images for better model generalization. Third, a hybrid visual explanation method is proposed by combining gradient-weighted class activation with local interpretable model agnostic explanation to generate a visual explanation using a heatmap. Fourth, an exhaustive performance analysis of the proposed approach on ultrasound images collected from three different Indian hospitals is presented to showcase its efficacy for computer-aided cholelithiasis diagnosis. Fifth, a team of radiologists evaluates and validates the prediction and respective visual explanations made using the proposed approach. The results reveal that the proposed cholelithiasis classification approach beats the performance of state-of-the-art pre-trained CNN and Vision Transformer models. The heatmap generated through the proposed hybrid explanation method offers detailed visual explanations to enhance transparency and trustworthiness in the medical domain.

Prenatal Diagnostics Using Deep Learning: A Dual Approach to Plane Localization and Cerebellum Segmentation in Ultrasound Images.

The fetal ultrasound examination is the significant task of mid-term pregnancy inspection and the accurate localization as well as the segmentation of the cerebellum is crucial for clinical diagnosis. This research focuses on developing deep learning techniques for prenatal prediction of neurodevelopmental disorders using 5th-month ultrasound brain images. The study introduces two specialized convolutional neural network (CNN) architectures: the differential CNN for plane localization and the dual CNN for cerebellum segmentation which are critical for accurate diagnostics during prenatal care. The differential CNN incorporates six different convolutional operators to capture diverse features for precise localization of specific planes within images. The dual CNN architecture integrates both the original image and complementary information such as feature maps, to enhance segmentation accuracy for the cerebellum. The models are trained on annotated datasets of ultrasound images, validated, and tested on separate datasets. The effectiveness of the proposed CNN architectures is determined by performing the specific tasks of plane localization and cerebellum segmentation, respectively. The proposed models achieved high performances of 98.6% and 0.956% from accuracy and dice coefficient (DSC) compared to existing approaches in medical image analysis. The findings of this study have significant implications for the prenatal prediction of neurodevelopmental disorders, offering a promising advancement in prenatal care and early diagnostics. The custom CNN architectures tailored to the specific tasks of plane localization and cerebellum segmentation highlight the importance of task-specific model design in medical imaging. While the study acknowledges certain limitations and challenges, the power of deep learning is analyzed for marking the benefit of healthcare and neurodevelopmental disorder prediction during pregnancy.

Deep learning for forensic age estimation using orthopantomograms in children, adolescents, and young adults.

Forensic age estimation from orthopantomograms (OPGs) can be performed more quickly and accurately using convolutional neural networks (CNNs), making them an ideal extension to standard forensic age estimation methods. This study evaluates improvements in forensic age prediction for children, adolescents, and young adults by training a custom CNN from a previous study, using a larger, diverse dataset with a focus on dental growth features. 21,814 OPGs from 13,766 individuals aged 1 to under 25 years were utilized. The custom CNN underwent 1000 epochs of training and validation using 16,000 and 4000 OPGs, respectively. The best model was chosen by the least mean absolute error (MAE) and evaluated with an additional test dataset of 1814 independent OPGs. Furthermore, the CNN was applied to OPGs from 15 available forensic age estimations conducted by experts certified by the Study Group on Forensic Age Diagnostics (AGFAD), and the results were compared. A MAE of 0.93 ± 0.81 years and a mean-signed error (MSE) of -0.06 ± 1.23 years were achieved in the test dataset. 63% of predictions were accurate within 1 year, and 95% within 2.5 years. Results of the CNN were comparable to those obtained by experts, effectively highlighting discrepancies in the reported ages of individuals. Using a large and diverse dataset along with custom deep learning techniques, forensic age estimation can be significantly improved, often providing predictions accurate to within 1 year. This approach offers a reliable, robust, and objective complement to standard forensic age estimation methods. Question The potential of custom convolutional neural networks for forensic age estimation, along with a large, diverse dataset, warrants further investigation, offering valuable support to experts. Findings For 1814 test-orthopantomograms, 63% of predictions were accurate within 1 year and 95% within 2.5 years, similar to expert estimates in 15 forensic cases. Clinical relevance Many individuals' fates depend on accurate age estimation. Forensic age estimation can benefit from applying CNN-based methods to further enhance reliability and accuracy.

Enhancing Facial Expression Recognition with Robust CNN Architectures and Adaptive Preprocessing Techniques

Facial expression recognition (FER) is an essential technology at the intersection of artificial intelligence (AI), computer vision, and psychology. This study proposes a novel framework for FER, aiming to improve system robustness and generalization, especially under variable real-world conditions. Using the FER2013 dataset, this research combines an adaptive preprocessing pipeline with a custom Convolutional Neural Network (CNN) architecture. Key preprocessing steps include normalization, rotation, and flipping to improve data quality and diversity. The CNN architecture combines regularization methods, including dropout and L2 regularization. Dynamic hyperparameter tuning and early stopping optimize performance and prevent overfitting. Normalized confusion matrix indicating strong recognition for well-represented emotions, such as happiness with 86% accuracy, and challenges with underrepresented categories like disgust. This research aims to contribute to the ongoing development of facial expression recognition systems by enhancing their robustness and generalization. While further refinement is needed, this work provides a step toward more accurate and adaptable FER models, with the potential to support advancements in human-computer interaction and various real-world applications.

Image-Based Detection and Classification of Malaria Parasites and Leukocytes with Quality Assessment of Romanowsky-Stained Blood Smears.

Malaria remains a global health concern, with 249 million cases and 608,000 deaths being reported by the WHO in 2022. Traditional diagnostic methods often struggle with inconsistent stain quality, lighting variations, and limited resources in endemic regions, making manual detection time-intensive and error-prone. This study introduces an automated system for analyzing Romanowsky-stained thick blood smears, focusing on image quality evaluation, leukocyte detection, and malaria parasite classification. Using a dataset of 1000 clinically diagnosed images, we applied feature extraction techniques, including histogram bins and texture analysis with the gray level co-occurrence matrix (GLCM), alongside support vector machines (SVMs), for image quality assessment. Leukocyte detection employed optimal thresholding segmentation utility (OTSU) thresholding, binary masking, and erosion, followed by the connected components algorithm. Parasite detection used high-intensity region selection and adaptive bounding boxes, followed by a custom convolutional neural network (CNN) for candidate identification. A second CNN classified parasites into trophozoites, schizonts, and gametocytes. The system achieved an F1-score of 95% for image quality evaluation, 88.92% for leukocyte detection, and 82.10% for parasite detection. The F1-score-a metric balancing precision (correctly identified positives) and recall (correctly detected instances out of actual positives)-is especially valuable for assessing models on imbalanced datasets. In parasite stage classification, CNN achieved F1-scores of 85% for trophozoites, 88% for schizonts, and 83% for gametocytes. This study introduces a robust and scalable automated system that addresses critical challenges in malaria diagnosis by integrating advanced image quality assessment and deep learning techniques for parasite detection and classification. This system's adaptability to low-resource settings underscores its potential to improve malaria diagnostics globally.

A Deep Learning Framework for Multi-Class Skin Cancer Diagnosis: Addressing Class Imbalance and Performance Optimization Through Convolutional Neural Network Architectures

Skin cancer, particularly melanoma, is one of the most fatal forms of cancer worldwide, necessitating early detection for effective treatment. The HAM10000 dataset is used as a standard for skin lesion analysis in this study to use a Convolutional Neural Network (CNN) to sort seven different types of skin cancer into groups. The dataset consists of over 10,000 labeled dermoscopic images representing various classes such as melanoma, basal cell carcinoma, and benign keratosis. The methodology includes preprocessing techniques like resizing, normalization, and augmentation to enhance the diversity of the data and address class imbalance. The Adam optimizer, with a learning rate of 0.0001, trained a custom CNN model with multiple convolutional layers, batch normalization, and ReLU activations. The model’s performance was evaluated across training, validation, and testing subsets, achieving a training accuracy of 98.17% and a validation accuracy of 99.60% after 10 epochs. Evaluation tools, such as a classification report and a confusion matrix, showed high accuracy for common classes like melanocytic nevi but showed mistakes in underrepresented classes like vascular lesions and dermatofibroma. Compared with state-of-the-art techniques like ResNet and EfficientNet, the proposed model demonstrates competitive performance with fewer computational resources. The study emphasizes the potential of deep learning in automating skin cancer detection while identifying challenges like class imbalance and the need for improved generalization. The results show how important advanced augmentation techniques, transfer learning, and ensemble approaches are for making models work better. This research contributes to the ongoing development of reliable, automated diagnostic tools, aiming to improve clinical outcomes and assist dermatologists in accurate and efficient skin cancer diagnosis.

A robust CNN Deep Learning and InceptionV3 model Techniques for Enhanced Skin Cancer Detection

Skin cancer, a dangerous form of cancer, originates from DNA damage that causes unchecked cell growth, resulting in a rapid rise in its occurrence. Earlier research has explored the application of computerised analysis for detecting malignancy in images of skin lesions. Nonetheless, challenges remain, including complications with light reflections, variations in colour, and the diverse shapes and sizes of lesions. This research explores how deep learning techniques can improve the detection of skin cancer. We developed and meticulously evaluated two models: a custom-built Convolutional Neural Network (CNN) and a tailored InceptionV3 model that was pre-trained and adapted for our needs. Our primary goal with the HAM10000 dataset was to execute binary classification, distinguishing between malignant melanoma and benign nevus. The custom Convolutional Neural Network (CNN), noted for its efficient design, achieved a precision rate of 91.80%, while the more complex adapted InceptionV3 model secured an impressive accuracy of 95.72%. The findings showcased here demonstrate the effectiveness of both tailored and pre-trained deep learning models in identifying skin cancer. This showcases their capability to enhance diagnostic accuracy and efficiency, even when encountering various resource and computing constraints. This study highlights how deep learning can significantly improve the accuracy of skin cancer diagnoses, ultimately leading to better outcomes for patients.

Advanced CNN-Based Classification and Segmentation for Enhanced Breast Cancer Ultrasound Imaging

Breast cancer (BC) is one of the primary causes of mortality in women globally. Thus, early and exact identification is critical for effective treatment. This work investigates deep learning, more especially convolutional neural networks (CNNs), to classify BC from ultrasound images. We worked with a collection of breast ultrasound images from 600 patients. Our approach included extensive image preprocessing techniques, such as enhancement and overlay methods, before training various deep learning models with particular reference to VGG16, VGG19, ResNet50, DenseNet121, EfficientNetB0, and custom CNNs. Our proposed model achieved a remarkable classification accuracy of 97%, significantly outperforming established models like EfficientNetB0, MobileNet, and Inceptionv3. This research demonstrates the ability of advanced CNNs, when paired with good preprocessing, to significantly enhance BC classification from ultrasound images. We further used Grad-CAM to make the model interpretable so we may see which parts of the images the CNNs focus on when making decisions.

Comparative Analysis of CNN Algorithms for Mushroom Classification with Proposed Lightweight CNN Model

The classification of mushroom species presents significant ecologic and health-related challenges; advancement in classification techniques is required to gain reliable identifications. This study aims to explain a methodology that was devised and evaluated in the development of a novel, lightweight Convolutional Neural Network (CNN) designed specifically for the task of mushroom classification. The paper provides a custom CNN model that is computationally cost-effective and capable of high-precision classification, fit for real-time usage. Hence, the proposed model was evaluated on this dataset of curated mushroom images with traditional classifiers and state-of-the-art CNN architectures, such as EfficientNet-B7, ResNet50, InceptionV3, and MobileNetV2. The custom model is depth-wise separations engineered in such a way that while they reduce the computational load, they don't compromise the effectiveness of the model. The custom model achieved a test score of 0.68, which is moderate compared to more established models such as EfficientNet-B7 or ResNet50. This approach helps the model function effectively even on platforms having low computational resources. A comprehensive evaluation reveals that a custom CNN has reasonable accuracy in the identification of different mushroom species vis-à-vis existing models, but also significantly lightens the classification process.

Real -Time Network Packet Classification Exploiting Computer Vision Architectures

The upcoming 6G and NextG networks underscore the necessity of sophisticated security methods based on Artificial Intelligence (AI) in order to detect malicious activity and adjust to new threats. Because computer vision techniques may be used to recognize complex patterns, their incorporation into the cybersecurity industry is a promising development. In this work, we present a computationally effective categorization technique that enforces the real-time conversion of packets into pictures by directly acting upon the raw packets gathered at base stations. The suggested solution's novel features include its lightweight implementation, which well satisfies the requirements of upcoming 6G networks, and its network edge operation, which permits early threat detection as near to the packet origin as feasible. We examine the efficacy of this methodology in terms of F1-score and prediction time by employing cutting-edge computer vision architectures and a customized Convolutional Neural Network (CNN) to tackle an intrusion detection task utilizing a substantial 5G dataset. The CNN design is superior than complicated models, as demonstrated by the results of experiments. The CNN consistently beats the other cutting-edge computer vision models over several packet window sizes N (i.e., 10, 50, and 100 packets), reaching very high F1-scores (0.99593, 0.99860, and 0.99895). A scalability investigation reveals a trade- off between the performance and scalability of CNN, with higher N values resulting in longer prediction times. However, the scalability of the other computer vision models is superior, allowing for an ideal model selection free of compromises

Comparative Analysis of Conventional CNN v’s ImageNet Pretrained ResNet in Medical Image Classification

Convolutional Neural Networks (CNNs) are the prevalent technology in computer vision and have become increasingly popular for medical imaging data classification and analysis. In this field, due to the scarcity of medical data, pretrained ResNets on ImageNet can be considered a suitable first approach. This paper examines the medical imaging classification accuracy of conventional basic custom CNNs compared to ImageNet pretrained ResNets on various medical datasets in an effort to give more information about the importance of medical data and its preprocessing techniques for disease studies. Microscope-extracted cytological images were examined along with chest X-rays, MRI brain scans, and melanoma photographs. The medical images were examined in various sets, class combinations, and resolutions. Augmented image datasets and asymmetrical training and validation splits among the classes were also examined. Models were developed after they were tested and fine-tuned with respect to their network size, parameter values and network methods, image resolution, size of dataset, multitude, and genre of class. Overfitting was also examined, and comparative studies regarding the computational cost of different models were performed. The models achieved high accuracy in image classification that varies depending on the dataset and can be easily incorporated in future over-the-internet medical decision-supporting (telemedicine) environments. In addition, it appeared that conventional basic custom CNN overperformed ImageNet pretrained ResNets. The obtained results indicate the importance of utilizing medical image data as a testbed for improvements in CNN classification performance and the possibility of using CNNs and data preprocessing techniques for disease studies.

Semantic Based Image Indexing using Deep Learning

This paper introduces a semantic-based image indexing system utilizing a custom Convolutional Neural Network (CNN) for feature extraction and semantic embedding techniques for understanding image content. Traditional image indexing methods rely heavily on low-level visual features, often resulting in inaccurate or irrelevant results. By leveraging deep learning, our proposed system bridges this gap, allowing high-level semantic features to guide indexing and retrieval. Tested on the CIFAR-10 dataset, our approach demonstrates a significant improvement in precision, recall, and overall retrieval performance, showcasing its potential in real-world applications

A Customized CNN Architecture with CLAHE for Multi-Stage Diabetic Retinopathy Classification

This paper presents a customized Convolutional Neural Network (CNN) architecture for multi-stage detection of Diabetic Retinopathy (DR), a leading cause of vision impairment and blindness. The proposed model incorporates advanced image enhancement techniques, particularly Contrast Limited Adaptive Histogram Equalization (CLAHE), to improve the visibility of critical retinal features associated with DR. By integrating CLAHE with a finely tuned CNN, the proposed approach significantly enhances accuracy and robustness, allowing for more precise detection across various stages of DR. The proposed model was evaluated against several state-of-the-art techniques, with the customized CNN alone achieving an overall accuracy of 97.69%. The addition of CLAHE further boosts the performance, achieving an accuracy of 99.69%, underscoring the effectiveness of combining CLAHE with CNN for automated DR detection. This approach provides an efficient, scalable, and highly accurate solution for early and multistage DR detection, which is crucial for timely intervention and prevention of vision loss.

Leveraging Edge Computing and Deep Learning for the Real-Time Identification of Bean Plant Pathologies

Leveraging Edge Computing and Deep Learning for the Real-Time Identification of Bean Plant Pathologies

Enhancing advanced cervical cell categorization with cluster-based intelligent systems by a novel integrated CNN approach with skip mechanisms and GAN-based augmentation

Cervical cancer is one of the biggest challenges in global health, thus it forms a critical need for early detection technologies that could improve patient prognosis and inform treatment decisions. This development in the form of an early detection mechanism increases the chances of successful treatment and survival, as early diagnosis promptly offers interventions that can dramatically reduce the rate of deaths attributed to this disease. Here, a customized Convolutional Neural Network (CNN) model is proposed for cervical cancerous cell detection. It includes three convolutional layers with increasing filter sizes and max-pooling layers, followed by dropout and dense layers for improved feature extraction and robust learning. By using ResNet models as inspiration, the model further innovates by incorporating skip connections into the CNN design. By enabling direct feature transmission from earlier to later layers, skip links enhance gradient flow and help preserve important spatial information. By boosting feature propagation, this integration increases the model’s ability to recognize minute patterns in cervical cell images, hence increasing classification accuracy. In our methodology, the SIPaKMeD dataset has been employed which contains 4049 cervical cell images that are arranged into five different categories. To address class imbalance, Generative Adversarial Networks (GANs) have been applied for data augmentation; that is, synthetic images have been created, that improve the diversity of the dataset and further enhance the robustness of the same. The present model is astonishingly accurate in classifying five cervical cell types: koilocytes, superficial-intermediate, parabasal, dyskeratotic, and metaplastic, thus significantly enhancing early detection and diagnosis of cervical cancer. The model gives an excellent performance because it has a validation accuracy of 99.11% and a training accuracy of 99.82%. It is a reliable model in the diagnosis of cervical cancerous cells because it ensures advancement in the computer-assisted cervical cancer detection system.

Consensus-based iterative meta-pseudo-labeling for deep semi-supervised learning

Consensus-based iterative meta-pseudo-labeling for deep semi-supervised learning

Rice Leaf Disease Classification—A Comparative Approach Using Convolutional Neural Network (CNN), Cascading Autoencoder with Attention Residual U-Net (CAAR-U-Net), and MobileNet-V2 Architectures

Classifying rice leaf diseases in agricultural technology helps to maintain crop health and to ensure a good yield. In this work, deep learning algorithms were, therefore, employed for the identification and classification of rice leaf diseases from images of crops in the field. The initial algorithmic phase involved image pre-processing of the crop images, using a bilateral filter to improve image quality. The effectiveness of this step was measured by using metrics like the Structural Similarity Index (SSIM) and the Peak Signal-to-Noise Ratio (PSNR). Following this, this work employed advanced neural network architectures for classification, including Cascading Autoencoder with Attention Residual U-Net (CAAR-U-Net), MobileNetV2, and Convolutional Neural Network (CNN). The proposed CNN model stood out, since it demonstrated exceptional performance in identifying rice leaf diseases, with test Accuracy of 98% and high Precision, Recall, and F1 scores. This result highlights that the proposed model is particularly well suited for rice leaf disease classification. The robustness of the proposed model was validated through k-fold cross-validation, confirming its generalizability and minimizing the risk of overfitting. This study not only focused on classifying rice leaf diseases but also has the potential to benefit farmers and the agricultural community greatly. This work highlights the advantages of custom CNN models for efficient and accurate rice leaf disease classification, paving the way for technology-driven advancements in farming practices.

A hybrid approach of vision transformers and CNNs for detection of ulcerative colitis

Ulcerative Colitis is an Inflammatory Bowel disease caused by a variety of factors that lead to a serious impact on the quality of life of the patients if left untreated. Due to complexities in the identification procedures of this disease, the treatment timeline and quality can be severely affected, leading to further consequences for the sufferer. The difficulties in identification are due to high patients to healthcare professionals ratio. Researchers have proposed variety of machine/deep learning methods for automated detection of ulcerative colitis, however, several challenges exists including class imbalance problem, comprehensive feature extraction and accurate classification. We propose a novel method for accurate detection of ulcerative colitis with augmentation techniques to overcome class imbalance issue, a comprehensive feature vector extraction using custom architecture of Vision Transformer (ViT) and accurate classification using customized Convolutional Neural Network (CNN). We used the TMC-UCM and LIMUC datasets in this research for training and testing of proposed method and achieved accuracy of 90% with AUC-ROC scores of 0.91, 0.81, 0.94, and 0.94 for the endoscopic classes of Mayo 0, Mayo 1, Mayo 2, and Mayo 3 respectively. We have compared the proposed method with existing state of the art methods and conclude that the proposed method outperforms the existing methods.

LesionNet: an automated approach for skin lesion classification using SIFT features with customized convolutional neural network.

Accurate detection of skin lesions through computer-aided diagnosis has emerged as a critical advancement in dermatology, addressing the inefficiencies and errors inherent in manual visual analysis. Despite the promise of automated diagnostic approaches, challenges such as image size variability, hair artifacts, color inconsistencies, ruler markers, low contrast, lesion dimension differences, and gel bubbles must be overcome. Researchers have made significant strides in binary classification problems, particularly in distinguishing melanocytic lesions from normal skin conditions. Leveraging the "MNIST HAM10000" dataset from the International Skin Image Collaboration, this study integrates Scale-Invariant Feature Transform (SIFT) features with a custom convolutional neural network model called LesionNet. The experimental results reveal the model's robustness, achieving an impressive accuracy of 99.28%. This high accuracy underscores the effectiveness of combining feature extraction techniques with advanced neural network models in enhancing the precision of skin lesion detection.

Hand Gesture Recognition Using Ultrasonic Array with Machine Learning

In the field of gesture recognition technology, accurately detecting human gestures is crucial. In this research, ultrasonic transducers were utilized for gesture recognition. Due to the wide beamwidth of ultrasonic transducers, it is difficult to effectively distinguish between multiple objects within a single beam. However, they are effective at accurately identifying individual objects. To leverage this characteristic of the ultrasonic transducer as an advantage, this research involved constructing an ultrasonic array. This array was created by arranging eight transmitting transducers in a circular formation and placing a single receiving transducer at the center. Through this, a wide beam area was formed extensively, enabling the measurement of unrestricted movement of a single hand in the X, Y, and Z axes. Hand gesture data were collected at distances of 10 cm, 30 cm, 50 cm, 70 cm, and 90 cm from the array. The collected data were trained and tested using a customized Convolutional Neural Network (CNN) model, demonstrating high accuracy on raw data, which is most suitable for immediate interaction with computers. The proposed system achieved over 98% accuracy.