No abstract

Representing videos by densely extracted local space-time features has recently become a popular approach for analysing actions. In this paper, we tackle the problem of categorising human actions by devising Bag of Words (BoW) models based on covariance matrices of spatio-temporal features, with the features formed from histograms of optical flow. Since covariance matrices form a special type of Riemannian manifold, the space of Symmetric Positive Definite (SPD) matrices, non-Euclidean geometry should be taken into account while discriminating between covariance matrices. To this end, we propose to embed SPD manifolds to Euclidean spaces via a diffeomorphism and extend the BoW approach to its Riemannian version. The proposed BoW approach takes into account the manifold geometry of SPD matrices during the generation of the codebook and histograms. Experiments on challenging human action datasets show that the proposed method obtains notable improvements in discrimination accuracy, in comparison to several state-of-the-art methods.

In this study, the capability of artificial neural networks and multiple linear regression methods for modeling the tensile properties of cotton-covered nylon core yarns based on process parameters were investigated. The developed models were assessed by verifying Mean Square Error (MSE) and Correlation Coefficient (R-value) the test data prediction. The results indicated that artificial neural network algorithm has better performance in comparison with multiple linear regression. The difference between the mean square error of predicting these two models for breaking strength and breaking elongation was 0.365 and 0.119, respectively. The five-fold cross-validation technique was used to evaluate the performance of artificial neural network algorithm. Moreover, the weight decay technique was also used for preventing the memorization.

Conventionally, weave repeat is identified manually by extracting individual warp or weft yarns from the fabric. This process can be troublesome and time-consuming. Therefore, automatic methods capable of identifying woven fabric repeat can be very useful.This paper describes the application of a new algorithm using image processing techniques for the development of an automatic method, capable of identifying weave repeat. This method is based on scanning and obtaining a gray scale image of the original sample and enhancing it by morphological operations. The enhanced image is filtered by steerable vertical filters and then segmented into blocks showing either a warp or a weft point. The blocked image is divided into specific sub images, followed by operating sum over their columns and forming a matrix from them. A primary and secondary threshold is then defined giving rise to the formation of the weave pattern in the form of black and white squares. To identify the weave repeat, a matrix, replacing the black and white squares of the weave pattern by zero and one is produced. Then the first repeating row and column are found, showing the start of the next repeat vertically and horizontally, leading to the identification of weave repeat.

No abstract