Image recognition (Python, Keras)

There is a small task, the results of which are going to be used for demo purposes as a part of non-technical research in a field: 1 - Import the images and prepare them for analysis (3 datasets are available); 2 - Resize the images; 3 - Feature extraction; 3.1. - Select 10 features (RGB mean, RGB variance, features derived from GLCM matrix and the ones selected); 4 - Standardization of the feature values; 5 - Make illustrations of the feature relationships (SOM, PCA); 6 - Try different classifiers, and optimize their parameters (K nearest neighbors, Regularized linear model with ridge regression, Multi-layer perceptron with 1 hidden layer and early stop committee (Output function Softmax)); 7 - Estimate the classifier accuracy; 8 - Prepare a report with the results; The whole technical task is available and would be shared.