Photo classification based on the presence of diagonal line using pre-trained DCNN VGG16

The diagonal line is one of the important linear components in a photograph. It contributes significantly towards improving the aesthetic perception and rendering high scores by enhancing the sense of direction. Automatic detection of diagonal lines in a photograph can link to numerous real-time applications like on-site guidance to amateur photographers by finding the presence of diagonal lines, aesthetic evaluation of photographs in various photography competitions, searching similar photographs containing diagonal lines from any database without manual labelling, and so on. A deep dive into the related literature reveals that very rare research has been conducted in the mentioned area. This paper presents a novel VGG16 Deep Convolutional Neural Network (DCNN) based approach to find diagonal lines in a photograph. The proposed approach classifies digital photographs into two categories: photographs containing diagonal lines or not by the application of transfer learning. The proposed model is implemented on the ground truth dataset of 5,683 images and the satisfactory results have been achieved. The contribution of the paper is significant due to the fact that the existence of a similar classifier for digital photographs is zero in the literature. The proposed work also contributes towards the generation of the diagonal line image dataset called Diagonal-line Containing Image (DCI) which will be useful for future research in the domain.