Litcius/Paper detail

Optical Character Recognition using Ensemble of SVM, MLP and Extra Trees Classifier

Lyagala Abhishek

20202020 International Conference for Emerging Technology (INCET)41 citationsDOI

Abstract

This paper deals with retrieval of contents of any printed or handwritten document. Maximally Stable Extremal Regions (MSER) algorithm along with region-growing methods are used for the detection of printed regions. Histogram of Oriented Gradients (HOG features) are used for feature extraction. Various machine learning algorithms, namely Decision Trees, Random Forest, Extra Trees Classifier, MLP, and SVM along with ensemble method were used for classification, and the accuracies compared.

Topics & Concepts

Pattern recognition (psychology)Artificial intelligenceSupport vector machineComputer scienceHistogramRandom forestHistogram of oriented gradientsClassifier (UML)Feature extractionDecision treeCharacter recognitionEnsemble learningRandom subspace methodImage (mathematics)Handwritten Text Recognition TechniquesVehicle License Plate RecognitionFace and Expression Recognition
Optical Character Recognition using Ensemble of SVM, MLP and Extra Trees Classifier | Litcius