Object classification on image using machine learning

Author: giorgi tetvadze
Keywords: Machine learning, Deep learning, Object classification
Annotation:

In recent years the big data has become easily accessible by general public. Growth in number and diversity of data available on the internet made possible to create machine learning applications. Machine learning is a subfield of artificial intelligence that can learn and improve its accuracy by learning over data without defining required properties explicitly. In current thesis our goal is to create deep learning models to identify damaged cars on colored images. To achieve desirable results we used widely popular and practically tested deep learning architectures and algorithms. Our final model showed 96% accuracy over validation data. Research can be made to further increase current achievements.