Automatic target image detection for morphing

• We design an algorithm to generate entire morphing process automatically.
• We find the best target animal face from the database considering human face as a source.
• Target detection results are compared with different feature based approaches.

In this paper, we propose a novel approach for automatic target image detection for morphing based on 3D textons and contrast. Given the source image of a human frontal face and training images with human and animal faces, our algorithm automatically finds the target image from the database of animal images. There are three major advantages of our approach: (1) it solves the problem of manual target selection as done by the researchers in morphing; (2) automatic target detection achieves smooth transition from source to destination image; and (3) control points are automatically detected for morphing because the algorithm detects the target based on the matching features. The experiments were conducted with images of six different animals viz. cheetah, lion, deer, red fox, snow leopard and rhesus monkey. For a subjective verification, a large number of human annotations are used.