Handcrafted vs non-handcrafted features for computer vision classification

[abstract]

This work investigates the possibility of exploiting trained deep Convolutional Neural Networks (CNN) as a generic feature extractor and how it is possible to mix such algorithms with more traditional hand-crafted features for developing a generic computer vision system. Such a system is a single structure that can be used for synthesizing a large number of different image classification tasks. Three structures are proposed for creating the generic computer vision system starting from handcrafted and non-handcrafter features: i) remapping the output layer of a trained CNN to classify a different problem using an SVM; ii) exploiting the output of the penultimate layer of a trained CNN as a feature vector to feed an SVM; iii) merging the output of some deep layers, applying a dimensionality reduction method, and using these features as the input to an SVM. The application of feature transform techniques to reduce the dimensionality of feature sets coming from deep layers represents one of the main contributions of this paper. Three approaches are used for the non-handcrafted features: deep transfer learning features based on convolutional neural networks (CNN), principal component analysis network (PCAN), and the compact binary descriptor (CBD). Regarding hand-crafted features, a wide variety of state-of-the-art algorithms are considered: Local Ternary Patterns, Local Phase Quantization, Rotation Invariant Co-occurrence Local Binary Patterns, Completed Local Binary Patterns, Rotated local binary pattern image, Globally Rotation Invariant Multi-scale Co-occurrence Local Binary Pattern, and several others. The computer vision system based on the proposed approach was tested on many different datasets, demonstrating the strong performance achieved by the proposed architecture. The Wilcoxon signed rank test is used to compare the different methods; furthermore, the independence of the different methods is studied using the Q-statistic. To facilitate replication of our experiments, the MATLAB source code will be available at (https://www.dei.unipd.it/node/2357 +Pattern Recognition and Ensemble Classifiers).

Keywords Deep learning; transfer learning, non-handcrafted features, texture descriptors; texture classification; ensemble of descriptors.