Search

BackgroundRecent studies use machine-learning techniques to detect parasites in microscopy images automatically. However, these tools are trained and tested in specific datasets. Indeed, even if over-fitting is avoided during the improvements of computer vision applications, large differences are expected. Differences might be related to settings of camera (exposure, white balance settings, etc) and different blood film slides preparation. Moreover, generative adversial networks offer new opportunities in microscopy: data homogenization, and increase of images in case of imbalanced or small sample size.MethodsTaking into consideration all those aspects, in this paper, we describe a more complete view including both detection and generating synthetic images: i) an automated detection used to detect malaria parasites on stained blood smear images using machine learning techniques testing several datasets. ii) investigate transfer learning and further testing in different unseen datasets having different staining, microscope, resolution, etc. iii) a generative approach to create synthetic images which can deceive experts.ResultsThe tested architecture achieved 0.98 and 0.95 area under the ROC curve in classifying images with respectively thin and thick smear. Moreover, the generated images proved to be very similar to the original and difficult to be distinguished by an expert microscopist, which identified correcly the real data for one dataset but had 50% misclassification for another dataset of images.ConclusionThe proposed deep-learning architecture performed well on a classification task for malaria parasites classification. The automated detection for malaria can help the technician to reduce their work and do not need any presence of experts. Moreover, generative networks can also be applied to blood smear images to generate useful images for microscopists. Opening new ways to data augmentation, translation and homogenization.

The relationship between structure and function is of interest in many research fields involving the study of complex biological processes. In neuroscience in particular, the fusion of structural and functional data can help understanding the underlying principles of the operational networks in the brain. To address this issue, this paper proposes a constrained autoregressive model leading to a representation of “effective” connectivity that can be used to better understand how the structure modulates the function. Or simply, it can be used to find novel biomarkers characterizing groups of subjects. In practice, an initial structural connectivity representation is re-weighted to explain the functional co-activations. This is obtained by minimizing the reconstruction error of an autoregressive model constrained by the structural connectivity prior. The model has been designed to also include indirect connections, allowing to split direct and indirect components in the functional connectivity, and it can be used with raw and deconvoluted BOLD signal.The derived representation of dependencies was compared to the well known dynamic causal model, giving results closer to known ground-truth. Further evaluation of the proposed effective network was performed on two typical tasks. In a first experiment the direct functional dependencies were tested on a community detection problem, where the brain was partitioned using the effective networks across multiple subjects. In a second experiment the model was validated in a case-control task, which aimed at differentiating healthy subjects from individuals with autism spectrum disorder. Results showed that using effective connectivity leads to clusters better describing the functional interactions in the community detection task, while maintaining the original structural organization, and obtaining a better discrimination in the case-control classification task.HighlightsA method to combine structural and functional connectivity by using autoregressive model is proposed.The autoregressive model is constrained by structural connectivity defining coefficients for Granger causality.The usefulness of the generated effective connections is tested on simulations, ground-truth default mode network experiments, a classification and clustering task.The method can be used for direct and indirect connections, and with raw and deconvoluted BOLD signal.

The analysis of the brain from a connectivity perspective is revealing novel insights into brain structure and function. Discovery is, however, hindered by the lack of prior knowledge used to make hypotheses. Additionally, exploratory data analysis is made complex by the high dimensionality of data. Indeed, to assess the effect of pathological states on brain networks, neuroscientists are often required to evaluate experimental effects in case-control studies, with hundreds of thousands of connections. In this paper, we propose an approach to identify the multivariate relationships in brain connections that characterize two distinct groups, hence permitting the investigators to immediately discover the subnetworks that contain information about the differences between experimental groups. In particular, we are interested in data discovery related to connectomics, where the connections that characterize differences between two groups of subjects are found. Nevertheless, those connections do not necessarily maximize the accuracy in classification since this does not guarantee reliable interpretation of specific differences between groups. In practice, our method exploits recent machine learning techniques employing sparsity to deal with weighted networks describing the whole-brain macro connectivity. We evaluated our technique on functional and structural connectomes from human and murine brain data. In our experiments, we automatically identified disease-relevant connections in datasets with supervised and unsupervised anatomy-driven parcellation approaches and by using high-dimensional datasets.

In ultrasound imaging, manual extraction of contours of fetal anatomic structures from echographic images have been found to be very challenging due to speckles and low contrast characteristic features. Contours extracted are therefore associated with variability of human observers. In this case, the contours that are extracted are not reproducible and hence not reliable. This challenge has called for the need to develop a method that can accurately segment the fetal anatomic structures. This will help to estimate and measure the contours of the structures of fetal bodies such as the head circumference, femur length, etc. Most recent methods are able to integrate global shape and appearance. The drawback to most of these methods is that, they are not able to handle localized appearance variations. They only rely on an assumption of Gaussian gray value distribution and also require initialization near the optimal solution. In this manuscript random forest is used to segment head contour in fetal ultrasound scans acquired in low-cost settings, such as acquisition performed in rural areas of low-income countries using low-cost portable machines.

Integration of functional and structural brain connectivity is a topic receiving growing attention in the research community. Their fusion can, in fact, shed new light on brain functions. Targeting this issue, the manuscript proposes a constrained autoregressive model allowing to generate an “effective” connectivity matrix that model the structural connectivity integrating the functional activity. In practice, an initial structural connectivity representation is altered according to functional data, by minimizing the reconstruction error of an autoregressive model constrained by the structural prior. The proposed model has been tested in a community detection framework, where the brain is partitioned using the effective network across multiple subjects. Results showed that using the effective connectivity the resulting clusters better describe the functional interactions of different regions while maintaining the structural organization.

The estimation of covariance matrices is a crucial step in several statistical tasks. Especially when using few samples of a high dimensional representation of shapes, the standard maximum likelihood estimation (ML) of the covariance matrix can be far from the truth, is often rank deficient, and may lead to unreliable results. In this paper, we discuss regularization by prior knowledge using maximum a posteriori (MAP) estimates. We compare ML to MAP using a number of priors and to Tikhonov regularization. We evaluate the covariance estimates on both synthetic and real data, and we analyze the estimates' influence on a missing-data reconstruction task, where high resolution vertebra and cartilage models are reconstructed from incomplete and lower dimensional representations. Our results demonstrate that our methods outperform the traditional ML method and Tikhonov regularization.

Unsupervised shape clustering can facilitate the labeling of objects present in images, especially when not all the samples are already identied by humans or some mistakes can be present. Some medical imaging tasks can present these diculties. For example, vertebra shapes annotated from lateral x-ray acquisitions are relevant for prognosis and diagnosis of fracture risk and osteoporosis, but their analysis can be cumbersome. In this paper, we propose a fracture prognosis framework able to outperforms other supervised estimations, based on simulated annealing clustering and non-linear dimensionality reduction. The cohort in exam is divided into a case and control group, in which the former sustained one incident lumbar fracture and the latter maintained skeletal integrity from baseline to follow-up. All subjects are fracture-free at baseline and from a subset of a larger epidemiological population, selected matching at baseline with respect to age, height, weight, spine BMD, physical activities and smoking habits; in this way the known risk factors are isolated and the only dierence left is their shapes variations. Each vertebra is represented by 6 points on its contour. A classier is trained to separate cases and controls at baseline. Our experiments show the possibility of using unsupervised clustering for overcoming misleading human classication in challenging tasks.

Functional and structural connectivity convey different information about the brain. The integration of these different approaches is receiving growing attention from the research community, as it can shed new light on brain functions. This manuscript proposes a constrained autoregressive model with different lag-orders generating an “effective” connectivity matrix which models the structural connectivity integrating the functional activity. Multiple orders are investigated to observe how different time dependencies influence the effective connectivity. The proposed approach alters an initial structural connectivity representation according to functional data, by minimizing the reconstruction error of an autoregressive model constrained by the structural prior. The model is further validated in a case-control experiment, which aims at differentiating healthy subject and young patients affected by autism spectrum disorder.

International audience; We introduce a new framework for the automated and un-supervised segmentation of Multiple Sclerosis lesions from multimodal Magnetic Resonance images. It relies on a voxel-wise approach to detect local white matter abnormalities, with an a-contrario analysis, which takes into account local information. First, a voxel-wise comparison of multimodal patient images to a set of controls is performed. Then, region-based probabilities are estimated using an a-contrario approach. Finally, correction for multiple testing is performed. Validation was undertaken on a multi-site clinical dataset of 53 MS patients with various number and volume of lesions. We showed that the proposed framework outperforms the widely used FDR-correction for this type of analysis, particularly for low lesion loads.