Towards intelligent photo composition-automatic detection of unintentional dissection lines in environmental portrait photos

Good photos result from careful attention to various elements such as shutter speed, aperture, light exposure and composition. Over the years, there have been interests to develop digital still cameras with “composition advising” functions. However, not many composition-related functions have been realized in current digital cameras. This study proposed an approach for automatic detection of unintentional dissection lines which often degrade the aesthetics of an environmental portrait photo. The algorithm includes modules for face detection, ROI estimation, morphology filtering, edge detection and straight line detection. Experimental evaluations conducted to verify the performance of the algorithm show that the detection rate and false detection rate of general dissection lines are 80.87% and 33.61%, respectively.