Article
Computer Science, Information Systems
D. Venugopal, T. Jayasankar, Mohamed Yacin Sikkandar, Mohamed Ibrahim Waly, Irina Pustokhina, Denis A. Pustokhin, K. Shankar
Summary: Data fusion poses a challenge in the healthcare sector, with Deep Learning being the preferred method for diagnosing conditions like Intracerebral Haemorrhage. The proposed FFEDL-ICH model combines handcrafted and deep features to outperform existing models, showing significant improvement in its performance. Further research is recommended to enhance the model's performance using learning rate scheduling techniques for Deep Neural Networks.
CMC-COMPUTERS MATERIALS & CONTINUA
(2021)
Article
Cell Biology
Aohan Liu, Yuchen Guo, Jinhao Lyu, Jing Xie, Feng Xu, Xin Lou, Jun-hai Yong, Qionghai Dai
Summary: This study presents a cross-modality learning framework called Cross-DL for detecting and localizing intracranial abnormalities in head CT scans by learning from free-text imaging reports. The framework includes a discretizer that automatically extracts abnormality types and locations from reports and trains an image analyzer using dynamic multi-instance learning. With a large-scale training set, Cross-DL achieves accurate performance in detecting and localizing abnormalities.
CELL REPORTS MEDICINE
(2023)
Article
Medicine, General & Internal
Yazan Qiblawey, Anas Tahir, Muhammad E. H. Chowdhury, Amith Khandakar, Serkan Kiranyaz, Tawsifur Rahman, Nabil Ibtehaz, Sakib Mahmud, Somaya Al Maadeed, Farayi Musharavati, Mohamed Arselene Ayari
Summary: The study proposed a cascaded system to detect and quantify COVID-19 infections from CT images, achieving high sensitivity and specificity, precise localization of infections, and the ability to discriminate between different severity levels of infection. The system showed excellent performance in experiments and could accurately localize infections of various shapes and sizes, even in small infection regions rarely considered in recent studies.
Article
Engineering, Multidisciplinary
Hasan Ulutas, M. Emin Sahin, Mucella Ozbay Karakus
Summary: This study presents a fully automated deep-learning method specifically designed for COVID-19 diagnosis and prognostic analysis on embedded systems. The proposed method achieves high accuracy and demonstrates feasibility through CT scan classification and evaluation. The results suggest that this method has potential in diagnosing COVID-19 and supporting radiologists.
ALEXANDRIA ENGINEERING JOURNAL
(2023)
Article
Engineering, Biomedical
Guyue Zhang, Kaixing Chen, Shangliang Xu, Po Chuan Cho, Yang Nan, Xin Zhou, Chuanfeng Lv, Changsheng Li, Guotong Xie
Summary: The paper introduces a novel strategy of generating artificial lesions on non-lesion CT images to produce additional labeled training examples, aiming to address the issue of insufficient training examples for training deep learning models.
COMPUTERIZED MEDICAL IMAGING AND GRAPHICS
(2021)
Article
Computer Science, Artificial Intelligence
Zhou Tao, Lu Huiling, Yang Zaoli, Qiu Shi, Huo Bingqiang, Dong Yali
Summary: This study proposes an ensemble deep learning model for rapid detection of COVID-19 from CT images, achieving higher accuracy and sensitivity compared to individual classifiers. This approach can better meet the requirements for rapid detection of the novel coronavirus disease COVID-19.
APPLIED SOFT COMPUTING
(2021)
Article
Automation & Control Systems
Aniello Castiglione, Pandi Vijayakumar, Michele Nappi, Saima Sadiq, Muhammad Umer
Summary: It is widely acknowledged that early disclosure of COVID-19 can limit its spread significantly. The ADECO-CNN model offers a highly accurate and precise method for classifying infected and not infected patients, showing superiority over other pretrained CNN-based models. Extensive analysis demonstrates the ADECO-CNN model's ability to classify CT images with exceptional accuracy, sensitivity, precision, and specificity.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2021)
Article
Engineering, Biomedical
Yin Gao, Jennifer Xiong, Chenyang Shen, Xun Jia
Summary: This study investigated the robustness properties of deep neural networks for a lung nodule classification problem based on CT images and proposed a solution to improve robustness by retraining the last layer. The results showed that the DNN-SVM model can enhance model robustness and reduce the susceptibility of 5% CT image cubes to noise.
PHYSICS IN MEDICINE AND BIOLOGY
(2021)
Article
Computer Science, Information Systems
Reza Majidpourkhoei, Mehdi Alilou, Kambiz Majidzadeh, Amin Babazadehsangar
Summary: The paper presents a framework for identifying pulmonary nodules in lung CT images and uses a convolutional neural network (CNN) approach to automatically extract features and classify suspicious regions. The proposed system achieves high accuracy and sensitivity, making it suitable for real-time medical image analysis after training and validation.
MULTIMEDIA TOOLS AND APPLICATIONS
(2021)
Article
Computer Science, Artificial Intelligence
Pavani Tripathi, Yasmeena Akhter, Mahapara Khurshid, Aditya Lakra, Rohit Keshari, Mayank Vatsa, Richa Singh
Summary: Cataract is a common global eye disease and researchers have proposed a novel algorithm using near-infrared eye images and deep learning technology for cataract detection, showing promising results.
COMPUTER VISION AND IMAGE UNDERSTANDING
(2022)
Article
Environmental Sciences
Ehsan Khoramshahi, Roope Naesi, Stefan Rua, Raquel A. A. Oliveira, Axel Paivansalo, Oiva Niemelainen, Markku Niskanen, Eija Honkavaara
Summary: This article explores the use of drone techniques to identify alien barleys in oat fields. By employing a machine learning approach and drone images, the study successfully detects and localizes barley plants, providing a useful method for modern grain production industries.
Article
Computer Science, Information Systems
Huseyin Yasar, Murat Ceylan
Summary: The study utilized deep learning methods to automatically classify lung CT images for early diagnosis of Covid-19. A 23-layer CNN architecture was designed as a classifier, with additional training and testing processes for Alexnet and Mobilenetv2 CNN architectures conducted.
MULTIMEDIA TOOLS AND APPLICATIONS
(2021)
Article
Computer Science, Interdisciplinary Applications
Miguel Lopez-Perez, Arne Schmidt, Yunan Wu, Rafael Molina, Aggelos K. Katsaggelos
Summary: This study presents a novel ICH detection model based on Multiple Instance Learning and Deep Gaussian Processes, which can be trained with scan-level annotations and achieves good results in experiments. The DGPMIL model demonstrates superior performance in multiple experiments and shows great potential for applications in medical image classification.
COMPUTER METHODS AND PROGRAMS IN BIOMEDICINE
(2022)
Article
Medicine, General & Internal
Muhammad Waseem Sabir, Muhammad Farhan, Nabil Sharaf Almalki, Mrim M. Alnfiai, Gabriel Avelino Sampedro
Summary: This study compares four vision transformers for their effectiveness in accurately detecting and classifying patients with Pulmonary Fibrosis and localizing abnormalities in CT scans. The optimized Vision Transformer (ViT) demonstrated superior performance in accuracy and loss minimization.
FRONTIERS IN MEDICINE
(2023)
Article
Biology
Kadir Yildirim, Pinar Gundogan Bozdag, Muhammed Talo, Ozal Yildirim, Murat Karabatak, U. Rajendra Acharya
Summary: In this study, an automated kidney stone detection method using deep learning technology achieved an accuracy of 96.82% in identifying kidney stones, even in small sizes. The use of computer-aided diagnosis systems as auxiliary tools in diagnosis was demonstrated, showing promise for clinical applications. Additionally, the study highlighted the potential of employing popular deep learning methods in addressing challenging issues in urology.
COMPUTERS IN BIOLOGY AND MEDICINE
(2021)