Text this: Scene Recognition through Visual Attention and Image Features: A Comparison between SIFT and SURF Approaches