analyzing large amount of logos collected from the interne

A logo is a mark composed of a graph or a combination of text and graph. Typical visual elements in a logo design such as layout, shape, color, composition, and typeface. The graphical mark can exhibit interesting properties by mixing the elements in creative ways. Most previous research regarding the role of visual elements in logo design are of a qualitative nature. In this paper, we propose to incorporate visual feature extraction and analysis algorithms commonly utilized in computer vision to compute proper index and investigate key visual elements in logo design, including complexity, balance, and repetition. After analyzing a large amount of logos collected from the internet, we find out that most logos are of low complexity, high balance and exhibit a minor degree of repetition. We hope that the results obtained in this research serve as a catalyst to motivate further efforts in applying computer vision methods to the area of aesthetics and design.