Analysis of smartphone model identification using digital images

This paper is focused on smartphone model identification using image features. A total of 64 image features - broadly categorized into colour features, wavelet features and image quality features - are extracted from high-resolution smartphone images. A binary-class turned to multiclass support vector machine (SVM) is used as the classifier. Experimental results based on 1800 images captured with 10 different smartphone/tablet devices are promising in correctly identifying source smartphone model. Image quality metrics and wavelet features are shown to contain the most useful device/model information compared to colour features. However, compared to colour features, quality and wavelet features are highly sensitive to simple image modifications. The combined set of colour, quality and wavelet features achieves the overall best identification accuracy.