Fig. 13From: Scene search based on the adapted triangular regions and soft clustering to improve the effectiveness of the visual-bag-of-words modelEffect on the MAP performance by varying dictionary size on the image benchmark of the 15-SceneBack to article page