Adaptive key SURF feature extraction and application in unmanned vehicle dynamic object recognition

A new method based on adaptive Hessian matrix threshold of finding key SRUF( speeded up robust features) features is proposed and is applied to an unmanned vehicle for its dynamic object recognition and guided navigation. First,the object recognition algorithm based on SURF feature matching for unmanned vehicle guided navigation is introduced. Then,the standard local invariant feature extraction algorithm SRUF is analyzed,the Hessian Metrix is especially discussed,and a method of adaptive Hessian threshold is proposed which is based on correct matching point pairs threshold feedback under a close loop frame. At last,different dynamic object recognition experiments under different weather light conditions are discussed. The experimental result shows that the key SURF feature abstract algorithm and the dynamic object recognition method can be used for unmanned vehicle systems.

Integrated statistical and machine learning analysis provides insight into key influencing symptoms for distinguishing early-onset type 2 diabetes

Background: Being able to predict with confidence the early onset of type 2 diabetes from a suite of signs and symptoms (features) displayed by potential sufferers is desirable to commence treatment promptly. Late or inconclusive diagnosis can result in more serious health consequences for sufferers and higher costs for health care services in the long run.Methods: A novel integrated methodology is proposed involving correlation, statistical analysis, machine learning, multi-K-fold cross-validation, and confusion matrices to provide a reliable classification of diabetes-positive and -negative individuals from a substantial suite of features. The method also identifies the relative influence of each feature on the diabetes diagnosis and highlights the most important ones. Ten statistical and machine learning methods are utilized to conduct the analysis.Results: A published data set involving 520 individuals (Sylthet Diabetes Hospital, Bangladesh) is modeled revealing that a support vector classifier generates the most accurate early-onset type 2 diabetes status predictions with just 11 misclassifications (2.1% error). Polydipsia and polyuria are among the most influential features, whereas obesity and age are assigned low weights by the prediction models.Conclusion: The proposed methodology can rapidly predict early-onset type 2 diabetes with high confidence while providing valuable insight into the key influential features involved in such predictions.

A 3D Human Liver Model of Nonalcoholic Steatohepatitis

Background and Aims: To better understand nonalcoholic steatohepatitis (NASH) disease progression and to evaluate drug targets and compound activity, we undertook the devel-opment of an in vitro 3D model to mimic liver architecture and the NASH environment. Methods:We have developed an in vitro preclinical 3D NASH model by coculturing primary hu-man hepatocytes, human stellate cells, liver endothelial cells and Kupffer cells embedded in a hydrogel of rat collagen on a 96-well plate. A NASH-like environment was induced by ad-dition of medium containing free fatty acids and tumor ne-crosis factor-α. This model was then characterized by biochemical, imaging and transcriptomics analyses. Results:We succeeded in defining suitable culture conditions to main-tain the 3D coculture for up to 10 days in vitro, with the lowest level of steatosis and reproducible low level of inflammation and fibrosis. NASH disease was induced with a custom me-dium mimicking NASH features. The cell model exhibited the key NASH disease phenotypes of hepatocyte injury, steatosis, inflammation, and fibrosis. Hepatocyte injury was highlighted by a decrease of CYP3A4 expression and activity, without loss of viability up to day 10. Moreover, the model was able to stimulate a stable inflammatory and early fibrotic environ-ment, with expression and secretion of several cytokines. A global gene expression analysis confirmed the NASH induc-tion. Conclusions:This is a new in vitro model of NASH dis-ease consisting of four human primary cell-types that exhibits most features of the disease. The 10-day cell viability and cost effectiveness of the model make it suitable for medium throughput drug screening and provide attractive avenues to better understand disease physiology and to identify and characterize new drug targets.