Wireless capsule endoscopy multiclass classification using three-dimensional deep convolutional neural network model.

BACKGROUND: Wireless capsule endoscopy (WCE) is a patient-friendly and non-invasive technology that scans the whole of the gastrointestinal tract, including difficult-to-access regions like the small bowel. Major drawback of this technology is that the visual inspection of a large number of video frames produced during each examination makes the physician diagnosis process tedious and prone to error. Several computer-aided diagnosis (CAD) systems, such as deep network models, have been developed for the automatic recognition of abnormalities in WCE frames. Nevertheless, most of these studies have only focused on spatial information within individual WCE frames, missing the crucial temporal data within consecutive frames. METHODS: In this article, an automatic multiclass classification system based on a three-dimensional deep convolutional neural network (3D-CNN) is proposed, which utilizes the spatiotemporal information to facilitate the WCE diagnosis process. The 3D-CNN model fed with a series of sequential WCE frames in contrast to the two-dimensional (2D) model, which exploits frames as independent ones. Moreover, the proposed 3D deep model is compared with some pre-trained networks. The proposed models are trained and evaluated with 29 subject WCE videos (14,691 frames before augmentation). The performance advantages of 3D-CNN over 2D-CNN and pre-trained networks are verified in terms of sensitivity, specificity, and accuracy. RESULTS: 3D-CNN outperforms the 2D technique in all evaluation metrics (sensitivity: 98.92 vs. 98.05, specificity: 99.50 vs. 86.94, accuracy: 99.20 vs. 92.60). In conclusion, a novel 3D-CNN model for lesion detection in WCE frames is proposed in this study. CONCLUSION: The results indicate the performance of 3D-CNN over 2D-CNN and some well-known pre-trained classifier networks. The proposed 3D-CNN model uses the rich temporal information in adjacent frames as well as spatial data to develop an accurate and efficient model.

Skin Melanoma Detection in Microscopic Images Using HMM-Based Asymmetric Analysis and Expectation Maximization.

Melanoma is one of the deadliest types of skin cancer with increasing incidence. The most definitive diagnosis method is the histopathological examination of the tissue sample. In this paper, a melanoma detection algorithm is proposed based on decision-level fusion and a Hidden Markov Model (HMM), whose parameters are optimized using Expectation Maximization (EM) and asymmetric analysis. The texture heterogeneity of the samples is determined using asymmetric analysis. A fusion-based HMM classifier trained using EM is introduced. For this purpose, a novel texture feature is extracted based on two local binary patterns, namely local difference pattern (LDP) and statistical histogram features of the microscopic image. Extensive experiments demonstrate that the proposed melanoma detection algorithm yields a total error of less than 0.04%.

Multi-scale structural rich-club organization of the brain in full-term newborns: a combined DWI and fMRI study.

Objective.Our understanding of early brain development is limited due to rapid changes in white matter pathways after birth. In this study, we introduced a multi-scale cross-modal approach to investigate the rich club (RC) organization and topology of the structural brain networks in 40 healthy neonates using diffusion-weighted imaging and resting-state fMRI data.Approach.A group independent component analysis was first performed to identify eight resting state networks (RSNs) used as functional modules. A groupwise whole-brain functional parcellation was also performed at five scales comprising 100-900 parcels. The distribution of RC nodes was then investigated within and between the RSNs. We further assessed the distribution of short and long-range RC, feeder and local connections across different parcellation scales.Main results.Sharing the scale-free characteristic of small-worldness, the neonatal structural brain networks exhibited an RC organization at different nodal scales (NSs). The subcortical, sensory-motor and default mode networks were found to be strongly involved in the RC organization of the structural brain networks, especially in the zones where the RSNs overlapped, with an average cross-scale proportion of 45.9%, 28.5% and 10.5%, respectively. A large proportion of the connector hubs were found to be RC members for the coarsest (73%) to finest (92%) NSs. Our results revealed a prominent involvement of cortico-subcortical and cortico-cerebellar white matter pathways in the RC organization of the neonatal brain. Regardless of the NS, the majority (more than 65.2%) of the inter-RSN connections were long distance RC or feeder with an average physical connection of 105.5 and 97.4 mm, respectively. Several key RC regions were identified, including the insula and cingulate gyri, middle and superior temporal gyri, hippocampus and parahippocampus, fusiform gyrus, precuneus, superior frontal and precentral gyri, calcarine fissure and lingual gyrus.Significance.Our results emphasize the importance of the multi-scale connectivity analysis in assessing the cross-scale reproducibility of the connectivity results concerning the global and local topological properties of the brain networks. Our findings may improve our understanding of the early brain development.

A deep learning approach for mitosis detection: Application in tumor proliferation prediction from whole slide images.

The tumor proliferation, which is correlated with tumor grade, is a crucial biomarker indicative of breast cancer patients' prognosis. The most commonly used method in predicting tumor proliferation speed is the counting of mitotic figures in Hematoxylin and Eosin (H&E) histological slides. Manual mitosis counting is known to suffer from reproducibility problems. This paper presents a fully automated system for tumor proliferation prediction from whole slide images via mitosis counting. First, by considering the epithelial tissue as mitosis activity regions, we build a deep-learning-based region of interest detection method to select the high mitosis activity regions from whole slide images. Second, we learned a set of deep neural networks to detect mitosis detection from selected areas. The proposed mitosis detection system is designed to effectively overcome the mitosis detection challenges by two novel deep preprocessing and two-step hard negative mining approaches. Third, we trained a Support Vector Machine (SVM) classifier to predict the final tumor proliferation score. The proposed method was evaluated on the dataset of the Tumor Proliferation Assessment Challenge (TUPAC16) and achieved a 73.81 % F-measure and 0.612 weighted kappa score, respectively, outperforming all previous approaches significantly. Experimental results demonstrate that the proposed system considerably improves the tumor proliferation prediction accuracy and provides a reliable automated tool to support health care make-decisions.

Fast T2 mapping using multi-echo spin-echo MRI: A linear order approach.

PURPOSE: Multi-echo spin-echo sequence is commonly used for T2 mapping. The estimated values using conventional exponential fit, however, are hampered by stimulated and indirect echoes leading to overestimation of T2 . Here, we present fast analysis of multi-echo spin-echo (FAMESE) as a novel approach to decrease the complexity of the search space, which leads to accelerated measurement of T2 . METHODS: We developed FAMESE based on mathematical analysis of the Bloch equations in which the search space dimension decreased to only one. Then, we tested it in both phantom and human brain. Bland-Altman plot was used to assess the agreement between the estimated T2 values from FAMESE and the ones estimated from single-echo spin-echo sequence. The reliability of FAMESE was assessed by intraclass correlation coefficients. In addition, we investigated the noise stability of the method in synthetic and experimental data. RESULTS: In both phantom and healthy participants, FAMESE provided accelerated and SNR-resistant T2 maps. The FAMESE had a very good agreement with the single-echo spin echo for the whole range of T2 values. The intraclass correlation coefficient values for FAMESE were excellent (ie, 0.9998 and 0.9860 < intraclass correlation coefficient < 0.9942 for the phantom and humans, respectively). CONCLUSION: Our developed method FAMESE could be considered as a candidate for rapid T2 mapping with a clinically feasible scan time.

An Automatic and Robust Decision Support System for Accurate Acute Leukemia Diagnosis from Blood Microscopic Images.

This paper proposes an automatic and robust decision support system for accurate acute leukemia diagnosis from blood microscopic images. It is a challenging issue to segment leukocytes under uneven imaging conditions since features of microscopic leukocyte images change in different laboratories. Therefore, this paper introduces an automatic robust method to segment leukocyte from blood microscopic images. The proposed robust segmentation technique was designed based on the fact that if background and erythrocytes could be removed from the blood microscopic image, the remainder area will indicate leukocyte candidate regions. A new set of features based on hematologist visual criteria for the recognition of malignant leukocytes in blood samples comprising shape, color, and LBP-based texture features are extracted. Two new ensemble classifiers are proposed for healthy and malignant leukocytes classification which each of them is highly effective in different levels of analysis. Experimental results demonstrate that the proposed approach effectively segments leukocytes from various types of blood microscopic images. The proposed method performs better than other available methods in terms of robustness and accuracy. The final accuracy rate achieved by the proposed method is 98.10% in cell level. To the best of our knowledge, the image level test for acute lymphoblastic leukemia (ALL) recognition was performed on the proposed system for the first time that achieves the best accuracy rate of 89.81%.

Maximized Inter-Class Weighted Mean for Fast and Accurate Mitosis Cells Detection in Breast Cancer Histopathology Images.

Based on the Nottingham criteria, the number of mitosis cells in histopathological slides is an important factor in diagnosis and grading of breast cancer. For manual grading of mitosis cells, histopathology slides of the tissue are examined by pathologists at 40× magnification for each patient. This task is very difficult and time-consuming even for experts. In this paper, a fully automated method is presented for accurate detection of mitosis cells in histopathology slide images. First a method based on maximum-likelihood is employed for segmentation and extraction of mitosis cell. Then a novel Maximized Inter-class Weighted Mean (MIWM) method is proposed that aims at reducing the number of extracted non-mitosis candidates that results in reducing the false positive mitosis detection rate. Finally, segmented candidates are classified into mitosis and non-mitosis classes by using a support vector machine (SVM) classifier. Experimental results demonstrate a significant improvement in accuracy of mitosis cells detection in different grades of breast cancer histopathological images.

Staging of Fatty Liver Diseases Based on Hierarchical Classification and Feature Fusion for Back-Scan-Converted Ultrasound Images.

Fatty liver disease is progressive and may not cause any symptoms at early stages. This disease is potentially fatal and can cause liver cancer in severe stages. Therefore, diagnosing and staging fatty liver disease in early stages is necessary. In this paper, a novel method is presented to classify normal and fatty liver, as well as discriminate three stages of fatty liver in ultrasound images. This study is performed with 129 subjects including 28 normal, 47 steatosis, 42 fibrosis, and 12 cirrhosis images. The proposed approach uses back-scan conversion of ultrasound sector images and is based on a hierarchical classification. The proposed algorithm is performed in two parts. The first part selects the optimum regions of interest from the focal zone of the back-scan-converted ultrasound images. In the second part, discrimination between normal and fatty liver is performed and then steatosis, fibrosis, and cirrhosis are classified in a hierarchical basis. The wavelet packet transform and gray-level co-occurrence matrix are used to obtain a number of statistical features. A support vector machine classifier is used to discriminate between normal and fatty liver, and stage fatty cases. The results of the proposed scheme clearly illustrate the efficiency of this system with overall accuracy of 94.91% and also specificity of more than 90%.

An improved method for liver diseases detection by ultrasound image analysis.

Ultrasound imaging is a popular and noninvasive tool frequently used in the diagnoses of liver diseases. A system to characterize normal, fatty and heterogeneous liver, using textural analysis of liver Ultrasound images, is proposed in this paper. The proposed approach is able to select the optimum regions of interest of the liver images. These optimum regions of interests are analyzed by two level wavelet packet transform to extract some statistical features, namely, median, standard deviation, and interquartile range. Discrimination between heterogeneous, fatty and normal livers is performed in a hierarchical approach in the classification stage. This stage, first, classifies focal and diffused livers and then distinguishes between fatty and normal ones. Support vector machine and k-nearest neighbor classifiers have been used to classify the images into three groups, and their performance is compared. The Support vector machine classifier outperformed the compared classifier, attaining an overall accuracy of 97.9%, with a sensitivity of 100%, 100% and 95.1% for the heterogeneous, fatty and normal class, respectively. The Acc obtained by the proposed computer-aided diagnostic system is quite promising and suggests that the proposed system can be used in a clinical environment to support radiologists and experts in liver diseases interpretation.

A Novel CAD System for Mitosis detection Using Histopathology Slide Images.

Histopathology slides are one of the most applicable resources for pathology studies. As observation of these kinds of slides even by skillful pathologists is a tedious and time-consuming activity, computerizing this procedure aids the experts to have faster analysis with more case studies per day. In this paper, an automatic mitosis detection system (AMDS) for breast cancer histopathological slide images is proposed. In the proposed AMDS, the general phases of an automatic image based analyzer are considered and in each phase, some special innovations are employed. In the pre-processing step to segment the input digital histopathology images more precisely, 2D anisotropic diffusion filters are applied to them. In the training segmentation phase, the histopathological slide images are segmented based on RGB contents of their pixels using maximum likelihood estimation. Then, the mitosis and non-mitosis candidates are processed and hence that their completed local binary patterns are extracted object-wise. For the classification phase, two subsequently non-linear support vector machine classifiers are trained pixel-wise and object-wise, respectively. For the evaluation of the proposed AMDS, some object and region based measures are employed. Having computed the evaluation criteria, our proposed method performs more efficient according to f-measure metric (70.94% for Aperio XT scanner images and 70.11% for Hamamatsu images) than the methods proposed by other participants at Mitos-ICPR2012 contest in breast cancer histopathological images. The experimental results show the higher performance of the proposed AMDS compared with other competitive systems proposed in Mitos-ICPR2012 contest.