Reducing Computational Costs of an Embedded Classifier to Determine Leather Quality

• Fausto Sampaio IFCE
• Lucas Costa da Silva IFCE
• Pedro Pedrosa Rebouças Filho IFCE
• Elias Teodoro Silva IFCE

Resumo

Embedded computer vision applications have been incorporated in industrial automation, improving quality and safety of processes. Such systems involve pattern classifiers for specific functions that, many times, demand high memory footprint and processing time. This work suggests a strategy to choose GLCM (Gray Level Co-occurrence Matrix) features for an SVM classifier that can reduce computer resources utilization while preserving high classifier accuracy. Experimental results show a computing time of 3.18s, with an accuracy of 90%, to classify images of 69 × 92 pixels. This result will permit to embed computer vision applications in low-cost platforms.

Palavras-chave: Feature extraction, Support vector machines, Correlation, Computer vision, Training, Pattern recognition, Embedded systems, Computer Vision, Embedded Applications, Machine Learning, GLCM, SVM