A comparative study on image-based snake identification using machine learning.

Automated snake image identification is important from different points of view, most importantly, snake bite management. Auto-identification of snake images might help the avoidance of venomous snakes and also providing better treatment for patients. In this study, for the first time, it's been attempted to compare the accuracy of a series of state-of-the-art machine learning methods, ranging from the holistic to neural network algorithms. The study is performed on six snake species in Lar National Park, Tehran Province, Iran. In this research, the holistic methods [k-nearest neighbors (kNN), support vector machine (SVM) and logistic regression (LR)] are used in combination with a dimension reduction approach [principle component analysis (PCA) and linear discriminant analysis (LDA)] as the feature extractor. In holistic methods (kNN, SVM, LR), the classifier in combination with PCA does not yield an accuracy of more than 50%, But the use of LDA to extract the important features significantly improves the performance of the classifier. A combination of LDA and SVM (kernel = 'rbf') is achieved to a test accuracy of 84%. Compared to holistic methods, convolutional neural networks show similar to better performance, and accuracy reaches 93.16% using MobileNetV2. Visualizing intermediate activation layers in VGG model reveals that just in deep activation layers, the color pattern and the shape of the snake contribute to the discrimination of snake species. This study presents MobileNetV2 as a powerful deep convolutional neural network algorithm for snake image classification that could be used even on mobile devices. This finding pave the road for generating mobile applications for snake image identification.

A Tensor-Based Framework for rs-fMRI Classification and Functional Connectivity Construction.

Recently, machine learning methods have gained lots of attention from researchers seeking to analyze brain images such as Resting-State Functional Magnetic Resonance Imaging (rs-fMRI) to obtain a deeper understanding of the brain and such related diseases, for example, Alzheimer's disease. Finding the common patterns caused by a brain disorder through analysis of the functional connectivity (FC) network along with discriminating brain diseases from normal controls have long been the two principal goals in studying rs-fMRI data. The majority of FC extraction methods calculate the FC matrix for each subject and then use simple techniques to combine them and obtain a general FC matrix. In addition, the state-of-the-art classification techniques for finding subjects with brain disorders also rely on calculating an FC for each subject, vectorizing, and feeding them to the classifier. Considering these problems and based on multi-dimensional nature of the data, we have come up with a novel tensor framework in which a general FC matrix is obtained without the need to construct an FC matrix for each sample. This framework also allows us to reduce the dimensionality and create a novel discriminant function that rather than using FCs works directly with each sample, avoids vectorization in any step, and uses the test data in the training process without forcing any prior knowledge of its label into the classifier. Extensive experiments using the ADNI dataset demonstrate that our proposed framework effectively boosts the fMRI classification performance and reveals novel connectivity patterns in Alzheimer's disease at its early stages.

A Novel Fast Tensor-Based Preconditioner for Image Restoration.

Image restoration is one of the main parts of image processing. Mathematically, this problem can be modeled as a large-scale structured ill-posed linear system. Ill-posedness of this problem results in low-convergence rate of iterative solvers. For speeding up the convergence, preconditioning usually is used. Despite the existing preconditioners for image restoration, which are constructed based on approximations of the blurring matrix, in this paper, we propose a novel preconditioner with a different viewpoint. Here, we show that image restoration problem can be modeled as a tensor contractive linear equation. This modeling enables us to propose a new preconditioner based on an approximation of the blurring tensor operator. Due to the particular structure of the blurring tensor for zero boundaries, we show that the truncated higher order singular value decomposition of the blurring tensor is obtained very fast and so could be used as a preconditioner. Experimental results confirm the efficiency of this new preconditioner in image restoration and its outperformance in comparison with the other well-known preconditioners.

A novel weighted support vector machine based on particle swarm optimization for gene selection and tumor classification.

We develop a detection model based on support vector machines (SVMs) and particle swarm optimization (PSO) for gene selection and tumor classification problems. The proposed model consists of two stages: first, the well-known minimum redundancy-maximum relevance (mRMR) method is applied to preselect genes that have the highest relevance with the target class and are maximally dissimilar to each other. Then, PSO is proposed to form a novel weighted SVM (WSVM) to classify samples. In this WSVM, PSO not only discards redundant genes, but also especially takes into account the degree of importance of each gene and assigns diverse weights to the different genes. We also use PSO to find appropriate kernel parameters since the choice of gene weights influences the optimal kernel parameters and vice versa. Experimental results show that the proposed mRMR-PSO-WSVM model achieves highest classification accuracy on two popular leukemia and colon gene expression datasets obtained from DNA microarrays. Therefore, we can conclude that our proposed method is very promising compared to the previously reported results.