13 April 2018 in Journal of Digital Imaging
MHMohammad Sadegh Helfroush
Mohammad Sadegh Helfroush
1 April 2021 in Artificial Intelligence in Medicine
1 September 2021 in CAAI Transactions on Intelligence Technology
Early diagnosis of breast cancer, the most common disease among women around the world, increases the chance of treatment and is highly important. Nuclear atypia grading in histopathological images plays an important role in the final diagnosis and grading of breast cancer. Grading images by pathologists is a time consuming and subjective task. Therefore, the existence of a computer‐aided system for nuclear atypia grading is very useful and necessary. In this study, two automatic systems for grading nuclear atypia in breast cancer histopathological images based on deep learning methods are proposed. A patch‐based approach is introduced due to the large size of the histopathological images and restriction of the training data. In the proposed system I, the most important patches in the image are detected first and then a three‐hidden‐layer convolutional neural network (CNN) is designed and trained for feature extraction and to classify the patches individually. The proposed system II is based on a combination of the CNN for feature extraction and a two‐layer Long short‐term memory (LSTM) network for classification. The LSTM network is utilised to consider all patches of an image simultaneously for image grading. The simulation results show the efficiency of the proposed systems for automatic nuclear atypia grading and outperform the current related studies in the literature.
4 March 2021 in IEEE Transactions on Computational Social Systems
Nowadays, social network plays an important role in human life. Besides all advantages of social networks, the dissemination of rumors becomes a major concern for users, so it is important to find a way to limit the spread of misinformation as much as possible. Influence blocking maximization (IBM) is the problem of finding k nodes in a social graph to minimize the spread of rumor source at the end of a propagation process. In this article, we propose a two-step method called influence blocking maximization using martingale (IBMM) to solve IBM problem under competitive independent cascade model (ICM) with both (1-1/e-ε)-approximation guarantee and practical runtime efficiency. In the proposed method, first we calculate the number of required samples using a set of estimation techniques based on martingale; and then we generate the samples and find top-k savior nodes. We perform extensive experiments on three real-world data sets and three rumor sets with different behaviors. We both experimentally and theoretically show that the effectiveness of IBMM is close to greedy. The results also show that IBMM is very fast, in particular, for a network with 265,214 nodes, 420,045 edges, and a set of 50 high influential nodes as rumor, when k = 50, l=1, and ϵ = 0.5 IBMM returns the solution within 3.5 s.
29 July 2021 in IEEE Transactions on Systems, Man, and Cybernetics: Systems
Online social networks (OSNs) are one of the most popular services available on the Internet. They are also considered as the primary source of news for many people. In spite of all the advantages of OSNs, dissemination of rumors has become a major concern. Influence blocking maximization (IBM) is the problem of determining savior users to limit the spread of misinformation as much as possible. Although the IBM problem has been widely studied, most of existing works neglect the effect of deadline and delay. In this article, we study Temporal IBM and propose a sampling-based approach (TIBM-M) with both theoretical guarantee and practical runtime efficiency. We perform extensive experiments on four real-world datasets and both experimentally and theoretically show that TIBM-M consistently matches the greedy algorithm in terms of effectiveness while significantly improves the efficiency. The results also indicate that TIBM-M has better blocking effect than other methods in all conditions.