Use a suitable feature extraction algorithm (choice of the algorithm depends on what you want to do) and it would give a feature vector of one row as you require. If in case the feature extraction algorithm gives you a matrix and you want it as row vector, then, convert the matrix to a row vector using the following matlab command:

yourvector = yourmatrix(:);

Hi Lucy,

There is no pre-defined generic feature vector. Depending on your task, you can use different feature extraction methods, e.g. Histogram of Oriented Gradients (HOG), Local Binary Patterns (LBP) and even pre-trained Convolutional Neural Networks (CNN). For MATLAB, you should find the following libraries useful:

- http://www.vlfeat.org

- http://www.vlfeat.org/matconvnet/

after extraction, youcan use diag function [diag(diag(a))] or this operator(:)