Article
Computer Science, Artificial Intelligence
Loris Nanni, Michelangelo Paci, Sheryl Brahnam, Alessandra Lumini
Summary: This study proposes a method of data augmentation using image transformations to generate training sets for multiple classifiers to form an ensemble. The results show that building ensembles on the data level by combining different data augmentation methods produces classifiers that compete with and often surpass the best approaches reported in the literature.
NEURAL COMPUTING & APPLICATIONS
(2022)
Article
Plant Sciences
Rong-Zhou Qiu, Shao-Ping Chen, Mei-Xiang Chi, Rong-Bo Wang, Ting Huang, Guo-Cheng Fan, Jian Zhao, Qi-Yong Weng
Summary: In this study, a deep learning-based automatic identification model for detecting citrus greening disease was developed. The model achieved high accuracy in recognizing the five symptoms of the disease and demonstrated good generalization performance under different imaging conditions. The model also showed better detection performance for experienced users and can be used as a preliminary screening tool. Researchers developed a user-friendly app called "HLBdetector" that allows farmers to quickly detect citrus greening disease using just a mobile phone without the need for expert guidance.
FRONTIERS IN PLANT SCIENCE
(2022)
Article
Computer Science, Artificial Intelligence
Yongquan Yang, Haijun Lv, Ning Chen, Yang Wu, Jiayi Zheng, Zhongxi Zheng
Summary: Ensembles of deep CNNs play a crucial role in ensemble learning for artificial intelligence applications, but the increasing complexity of deep CNN architectures and large data dimensionality have made their usage costly. A new approach is proposed to find multiple models converging to local minima in the subparameter space of deep CNNs, which can improve generalization while being more affordable during training and testing stages.
PATTERN RECOGNITION
(2021)
Article
Computer Science, Artificial Intelligence
Md Sakib Ullah Sourav, Huidong Wang
Summary: In this study, an intelligent model based on transfer learning and deep convolutional neural networks was developed for fast and accurate identification of jute pests. The model showed good performance in classifying jute pests and has been integrated into mobile applications for practical use.
NEURAL PROCESSING LETTERS
(2023)
Article
Biology
Behrouz Rostami, D. M. Anisuzzaman, Chuanbo Wang, Sandeep Gopalakrishnan, Jeffrey Niezgoda, Zeyun Yu
Summary: Wound classification plays a crucial role in medical diagnosis, and a high-performance classifier can reduce time and financial costs. This study introduced a deep learning-based classifier with superior classification accuracy, showing potential applications in wound image classification.
COMPUTERS IN BIOLOGY AND MEDICINE
(2021)
Article
Computer Science, Artificial Intelligence
Tianyu Ma, Alan Q. Wang, Adrian V. Dalca, Mert R. Sabuncu
Summary: The convolutional neural network (CNN) is a commonly used architecture for computer vision tasks. A new building block called hyper-convolution is presented in this paper, which encodes the convolutional kernel using spatial coordinates and enables a more flexible architecture design. Experimental results showed that replacing regular convolutions with hyper-convolutions improved performance with fewer parameters and increased robustness against noise.
MEDICAL IMAGE ANALYSIS
(2023)
Article
Computer Science, Artificial Intelligence
Bruno Antonio, Davide Moroni, Massimo Martinelli
Summary: In recent times, there has been a trend in computer vision to achieve minor improvements compared to considerable increases in complexity. To reverse this trend, the authors propose a novel method to improve image classification performances without increasing complexity, by revisiting ensembling and making efficient adaptive ensembles.
Article
Engineering, Biomedical
Abderazzak Ammar, Omar Bouattane, Mohamed Youssfi
Summary: Cardiac cine MRI, using fully convolutional neural networks, can accurately segment heart structures and predict diseases. An automated pipeline for heart segmentation and diagnosis was proposed, achieving nearly state-of-the-art accuracy in the segmentation contest and disease classification challenge of the ACDC challenge.
COMPUTERIZED MEDICAL IMAGING AND GRAPHICS
(2021)
Article
Engineering, Biomedical
Abderazzak Ammar, Omar Bouattane, Mohamed Youssfi
Summary: This study proposes an automated pipeline for cardiac segmentation and diagnosis using MRI images and deep learning techniques. By combining three classifiers, the system achieved a high accuracy for heart disease classification on unseen data.
COMPUTERIZED MEDICAL IMAGING AND GRAPHICS
(2021)
Article
Engineering, Biomedical
Abderazzak Ammar, Omar Bouattane, Mohamed Youssfi
Summary: This study utilizes cardiac cine magnetic resonance imaging for cardiac structure segmentation and disease prediction, achieving impressive results in medical imaging competitions. The automated pipeline proposed in the research uses deep learning for cardiac structure segmentation and disease diagnosis.
COMPUTERIZED MEDICAL IMAGING AND GRAPHICS
(2021)
Article
Engineering, Biomedical
Abderazzak Ammar, Omar Bouattane, Mohamed Youssfi
Summary: This study proposes an automated pipeline for cardiac MRI segmentation and diagnosis using fully convolutional neural networks, achieving nearly state-of-the-art accuracy for both segmentation and disease classification challenges.
COMPUTERIZED MEDICAL IMAGING AND GRAPHICS
(2021)
Article
Biology
Eduardo Aguilar, Bhalaji Nagarajan, Petia Radeva
Summary: Deep learning is a machine learning technique that has revolutionized research with its impressive results. Using ensembles of Convolutional Neural Networks (CNN) can achieve high robustness and accuracy in computer vision challenges. Choosing the right number of models and selecting the best models from a trained pool is crucial. Random selection or exhaustive search methods have limitations. Employing an uncertainty-aware epistemic method can significantly improve performance and reduce computing costs.
COMPUTERS IN BIOLOGY AND MEDICINE
(2022)
Article
Computer Science, Information Systems
Jae-Min Lee, Min-Seok Seo, Dae-Han Kim, Sang-Woo Lee, Jong-Chan Park, Dong-Geol Choi
Summary: In this work, the authors propose a Split-and-Share Module (SSM), which splits given features into parts and shares them among multiple sub-classifiers, in order to improve the performance of image classification tasks and identify structural characteristics within the features. SSM can be easily integrated into various architectures and has been validated to show significant improvements over baseline architectures.
Article
Biology
Eduardo Aguilar, Bhalaji Nagarajan, Petia Radeva
Summary: Deep learning is a powerful machine learning technique that has produced impressive results in various real-life problems. Ensemble models of Convolutional Neural Networks (CNN) have proven to be robust and accurate in computer vision challenges. The selection of models in an ensemble is crucial, and uncertainty-aware methods may help determine the best groups of CNN models.
COMPUTERS IN BIOLOGY AND MEDICINE
(2022)
Article
Computer Science, Information Systems
Khaled Almezhghwi, Sertan Serte, Fadi Al-Turjman
Summary: The study proposes two artificial intelligence approaches utilizing deep learning for the classification of chest X-ray images. These methods, based on the AlexNet model and VGGNet16 method, can accurately identify lung diseases.
MULTIMEDIA TOOLS AND APPLICATIONS
(2021)
Article
Engineering, Biomedical
Ehsan Kozegar, Mohsen Soryani
BIOMEDICAL ENGINEERING-BIOMEDIZINISCHE TECHNIK
(2019)
Article
Computer Science, Interdisciplinary Applications
E. Kozegar, M. Soryani, H. Behnam, M. Salamati, T. Tan
IEEE TRANSACTIONS ON MEDICAL IMAGING
(2018)
Article
Computer Science, Artificial Intelligence
Ehsan Kozegar, Mohsen Soryani, Hamid Behnam, Masoumeh Salamati, Tao Tan
ARTIFICIAL INTELLIGENCE REVIEW
(2020)
Article
Food Science & Technology
Mohammad Reza Larijani, Ezzatollah Askari Asli-Ardeh, Ehsan Kozegar, Reyhaneh Loni
FOOD SCIENCE & NUTRITION
(2019)
Article
Acoustics
Amin Malekmohammadi, Sepideh Barekatrezaei, Ehsan Kozegar, Mohsen Soryani
Summary: Early detection of breast cancer symptoms is crucial in reducing mortality rates. The 3-D Automated Breast Ultrasound (ABUS) is popular for breast screening due to its sensitivity and reproducibility. However, manual evaluation of ABUS slices is challenging and time-consuming. To aid radiologists, a convolutional BiLSTM network was proposed for slice classification, which also identifies the approximate location of masses as a heat map. Results showed an 84% precision, 84% recall, 93% accuracy, 84% F1-score, and 97% AUC for the proposed model. FROC analysis revealed a sensitivity of 82% with two false positives per volume.
Article
Medicine, General & Internal
Xihe Kuang, Xiayu Xu, Leyuan Fang, Ehsan Kozegar, Huachao Chen, Yue Sun, Fan Huang, Tao Tan
Summary: Retinal images are important for diagnosing diseases. Segmentation of retinal vessels is crucial for the analysis of retinal images. Current methods focus on overall vessel structures and neglect small vessels. This paper proposes a method (UN-LPCOS) that combines unsupervised methods with a deep learning network to segment small retinal vessels effectively. A new metric (Se-sv) is also introduced to evaluate the segmentation performance. The proposed strategy achieves outstanding results on both overall vessel structure and small vessels.
FRONTIERS IN MEDICINE
(2023)
Article
Computer Science, Information Systems
Seyyed Mohammad Reza Hashemi, Hamid Hassanpour, Ehsan Kozegar, Tao Tan
Summary: An intelligent method for classifying cystoscopy images of bladder using a pre-trained CNN, PCA, LDA, and an ensemble classifier achieved an accuracy of 69.02 +/- 0.19 on 720 collected images, outperforming other methods.
Article
Engineering, Multidisciplinary
A. Ghanbari Sorkhi, S. M. R. Hashemi, H. Yarmohammadi, M. Iranpour Mobarakeh
Summary: Identification of drug-target protein interactions is crucial in drug discovery, and a novel method based on known interactions and similarity graphs was proposed in this paper. The method utilized WNNM to detect interactions and showed improved performance on benchmark datasets across various criteria like AUC and AUPR.
INTERNATIONAL JOURNAL OF ENGINEERING
(2021)
Article
Mathematics
Ehsan Kozegar
Summary: Over the last three decades, artificial intelligence in medical diagnosis tasks has attracted much attention. This paper proposes a new computer aided diagnosis system to classify four types of cystoscopic images, achieving an accuracy of 63% which outperforms base VGG-Nets and other competing methods.
INTERNATIONAL JOURNAL OF NONLINEAR ANALYSIS AND APPLICATIONS
(2021)
Article
Mathematics
S. M. R. Hashemi, H. Hassanpour, E. Kozegar, T. Tan
INTERNATIONAL JOURNAL OF NONLINEAR ANALYSIS AND APPLICATIONS
(2020)
Proceedings Paper
Computer Science, Software Engineering
Javad Ghofrani, Ehsan Kozegar, Anna Lena Fehlhaber, Mohammad Divband Soorati
SPLC'19: PROCEEDINGS OF THE 23RD INTERNATIONAL SYSTEMS AND SOFTWARE PRODUCT LINE CONFERENCE, VOL A
(2020)
Article
Mathematics
Seyyed Mohammadreza Hashemi, Hamid Hassanpour, Ehsan Kozegar, Tao Tan
INTERNATIONAL JOURNAL OF NONLINEAR ANALYSIS AND APPLICATIONS
(2019)
Proceedings Paper
Computer Science, Interdisciplinary Applications
Seyyed Mohammad. R. Hashemi, Ehsan Kozegar, Mohammad Mahdi Deramgozin, Behrouz Minaei-Bidgoli
2019 IEEE 5TH CONFERENCE ON KNOWLEDGE BASED ENGINEERING AND INNOVATION (KBEI 2019)
(2019)