“Emerging Images Synthesis from Photographs” by Jian, Chu, Lee, Ku, Wang, et al. …
Conference:
Type(s):
Entry Number: 39
Title:
- Emerging Images Synthesis from Photographs
Presenter(s)/Author(s):
Abstract:
Emergence refers to a phenomenon by which human perceives complete objects in a seemingly noisy image not by recognizing local parts of image but viewing the image as a whole. The Dalmatian dog image created by R. C. James is probably the best demonstration of emergence [Bach 2002]. It shows that local windows from the image reveal nothing but meaningless, complex and random black splats. Only when the image is viewed as a whole, a Dalmatian dog suddenly appears. The absence of meaningful information in local image parts largely hinders existing computer vision algorithms from recognizing emerging figures. Therefore, it makes the emergence an new type of CAPTCHA to tell human and machine apart.
References:
1. Achanta, R., Shaji, A., Smith, K., Lucchi, A., Fua, P., and Susstrunk, S. 2012. Slic superpixels compared to state-of-the-art superpixel methods. IEEE Trans. PAMI 34, 11, 2274–2282.
2. Bach, M., 2002. Hidden Figures — Dalmatian Dog.
3. Cheng, Ming-Ming, Z. G.-X. M. N. J. H. X. H. S.-M. 2011. Global contrast based salient region detection. In IEEE CVPR.
4. Mitra, N. J., Chu, H.-K., Lee, T.-Y., Wolf, L., Yeshurun, H., and Cohen-Or, D. 2009. Emerging images. ACM Trans. Graph. (Proc. SIGGRAPH Asia) 28, 5, 163:1–163:8.
