Mobile-Based Painting Photo Retrieval Using Combined Features

In paintings or artworks, sharing a photo of a painting using mobile phone is simple and fast. However, searching for information about specific captured photo of an unknown painting takes time and is not easy. No previous developments were introduced in the content-based indexing and retrieval (CBIR) field to ease the inconvenience of knowing the name and other information about an unknown painting through capturing photos by mobile phones. This work introduces an image retrieval framework on art paintings using shape, texture and color properties. With existing state-of-the-art developments, the proposed framework focuses on utilizing a feature combination of: generic Fourier descriptors (GFD), local binary patterns (LBP), Gray-level co-occurrence matrix (GLCM), and HSV histograms. After that, Locality Sensitive Hashing (LSH) method is used for image indexing and retrieval of paintings. The results are validated over a public database of seven different categories.