Robust and fast solutions for anatomical object detection and segmentation support the entire clinical workflow from diagnosis, patient stratification, therapy planning, intervention and follow-up. Current state-of-the-art techniques for parsing volumetric medical image data are typically based on machine learning methods that exploit large annotated image databases. Two main challenges need to be addressed, these are the efficiency in scanning high-dimensional parametric spaces and the need for representative image features which require significant efforts of manual engineering. We propose a pipeline for object detection and segmentation in the context of volumetric image parsing, solving a two-step learning problem: anatomical pose estimation and boundary delineation. For this task we introduce Marginal Space Deep Learning (MSDL), a novel framework exploiting both the strengths of efficient object parametrization in hierarchical marginal spaces and the automated feature design of Deep Learning (DL) network architectures. In the 3D context, the application of deep learning systems is limited by the very high complexity of the parametrization. More specifically 9 parameters are necessary to describe a restricted affine transformation in 3D, resulting in a prohibitive amount of billions of scanning hypotheses. The mechanism of marginal space learning provides excellent run-time performance by learning classifiers in clustered, high-probability regions in spaces of gradually increasing dimensionality. To further increase computational efficiency and robustness, in our system we learn sparse adaptive data sampling patterns that automatically capture the structure of the input. Given the object localization, we propose a DL-based active shape model to estimate the non-rigid object boundary. Experimental results are presented on the aortic valve in ultrasound using an extensive dataset of 2891 volumes from 869 patients, showing significant improvements of up to 45.2% over the state-of-the-art. To our knowledge, this is the first successful demonstration of the DL potential to detection and segmentation in full 3D data with parametrized representations.
Although a majority of somatic mutations in cancer are passengers, their mutational signatures provide mechanistic insights into mutagenesis and DNA repair processes. Mutational signature SBS8 is common in most cancers, but its etiology is debated. Incorporating genomic, epigenomic, and cellular process features for multiple cell-types we develop genome-wide composite epigenomic context-maps relevant for mutagenesis and DNA repair. Analyzing somatic mutation data from multiple cancer types in their epigenomic contexts, we show that SBS8 preferentially occurs in gene-poor, lamina-proximal, late replicating heterochromatin domains. While SBS8 is uncommon among mutations in non-malignant tissues, in tumor genomes its proportions increase with replication timing and speed, and checkpoint defects further promote this signature-suggesting that SBS8 probably arises due to uncorrected late replication errors during cancer progression. Our observations offer a potential reconciliation among different perspectives in the debate about the etiology of SBS8 and its relationship with other mutational signatures.
Handwriting analysis is a method to predict personality of an author and to better understand the writer. Allograph and allograph combination analysis is a scientific method of writer identification and evaluating the behavior. To make this computerized we considered six main different types of features: (i) size of letters, (ii) slant of letters and words, (iii) baseline, (iv) pen pressure, (v) spacing between letters and (vi) spacing between words in a document to identify the personality of the writer. Segmentation is used to calculate the features from digital handwriting and is trained to SVM which outputs the behavior of the writer. For this experiment 100 different writers were used for different handwriting data samples. The proposed method gives about 94% of accuracy rate with RBF kernel. In this paper an automatic method has been proposed to predict the psychological personality of the writer. The system performance is measured under two different conditions with the same sample.
Abstract:The Nash model was used for application of the Kalman ®lter. The state vector of the rainfall±runo system was constituted by the IUH (instantaneous unit hydrograph) estimated by the Nash model and the runo estimated by the Nash model using the Kalman ®lter. The initial values of the state vector were assumed as the average of 10% of the IUH peak values and the initial runo estimated from the average IUH. The Nash model using the Kalman ®lter with a recursive algorithm accurately predicted runo from a basin in Korea. The ®lter allowed the IUH to vary in time, increased the accuracy of the Nash model and reduced physical uncertainty of the rainfall±runo process in the river basin. #
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.