Combination of the two methods

For advanced comparison, we put the morphing process in one image. In this image, we put more than two objects in it. Then, we slice the image according to the objects. The following figure will show it.

Figure4-7. An example for the original input image

the original above image will be sliced into the below three images.

Figure4-8. (a) Figure4-8. (b) Figure4-8. (c) Figure4-8. (a) (b) (c) are the decompose by the Figure4-7

After we sliced the image, we can choose one method which is fitted the object property manually. By following the rules we have mentioned about, we can classify the first and second images (cloud) to blending and third image to shape-based

interpolation. The left up one is the source, and the right down one is target. In this example, we choose two clouds as one image to morphing with blending, and the tank morphing with shape-based interpolation.

Figure4-9. (a) Figure4-9. (b) Figure4-9. (c)

Figure4-9. (d) Figure4-9. (e) Figure4-9. (f) Figure4-9 (a) to (f) presents the image processing combined with two methods

The choice is not unique indeed. We can choose the left cloud as one object, the right cloud as the other object, needless to say, the tank as the third object. The

different choices make different results.

Figure4-10. (a) Figure4-10. (b) Figure4-10. (c)

Figure4-10. (d) Figure4-10. (e) Figure4-10. (f) Figure4-6 (a) to (f) presents the image processing combined with two methods by another choice

Chapter 5

Conclusions

5.1 Discussion

In this paper, our purpose is to compare the methods of morphological interpolations and level set methods - blending in morphing. In morphological interpolations, we have try to carry out shape-based interpolation, Hausdroff distance interpolation, distance-based interpolation, and median interpolation. We observe the in-between images in morphing process. The purpose of morphing is to construct a natural, smooth image sequence. To achieve this goal, we should confirm which effect we want at first. As described above, the results show the two methods cause completely different outcomes.

Mathematical interpolations are easy to understand and implement. A morphological morphing algorithm considers an image as a set, and the morphing transformation depends on operating two sets into the other one. In shape-based interpolation, it just uses two basic algebra operations – dilation and erosion, the in-between images can be obtained by idempotency of the morphed sets from the two side extremities. In Hausdroff distance interpolation, it just achieved by an operation – dilation. In distance-based interpolation, the morphed images are controlled by the distance functions. In median interpolation, the in-between images are just composed by influence zones.

On the contrary, level set methods are more complex. Level set methods based

we use image blending to achieve image morphing. Image blending is a low-level process to construct a set of image transition. The most important thing about using level set methods to do morphing is its ability of avoiding ghosting. The other additional advantage is that it is not restricted on empty intersection. By the way, in our experiment, we can find it spends more times than morphological shape-based interpolation does.

5.2 Future work

In this paper, we describe the detail of theory and application of mathematical morphology in chapter 2. Then, we discuss the detail of theory and application of level set methods in chapter 3. And, we present all the experimental results in chapter 4. But it still has several aspect should be enhanced.

Generally speaking, mathematical morphological interpolations cut the edge sharp. Besides, for empty intersection, except for Hausdroff distance interpolation is practicable, we just set the area by calculating the minimized distance between two original images.

In blending, unlike mathematical morphology, even though the source image and the target image are two quite different colors, the grayscale of pixels in morphed color image show in a more true form. Nevertheless, the whole morphed image stares blurred. Here, we try to clear the morphed image by using median interpolation morphed image to intensify the correspondent location pixels in blending morphed image in chapter 4 section 2. The impression on the in-between image are tolerable but could be advanced.

Collecting those disadvantages above all, we can do enhancement in two ways:

◆ For mathematical morphology:

Find some extra skill to make sure the grayscale of pixels in morphed image appear more real and may try another approach to solve empty intersection

◆ For blending method:

Find some image enhancement to strengthen the grayscale, let the image clear

References :

[1] T. Beier and S. Neely, “Feature-based image metamorphosis”. Computer Graphics (Proc. SIGGRAPH ’92), Vol. 26, No.2, PP.35–42, 1992.

[2] G. Wolberg, “Recent Advances in Image Morphing”. Computer Graphics International, 1996. Proceedings24-28 June 1996 Page(s):64 – 71, 1996 [3] G.Wolberg, “Digital ImageWarping”. IEEE Computer Society Press, Los

Alamitos, CA, 1990.

[4] P. Litwinowicz and L. Williams, “Animating images with drawings”. Computer Graphics (Proc. SIGGRAPH ’94), pages 409–412, 1994.

[5] S.-Y. Lee, K.-Y. Chwa, J. Hahn, and S. Y. Shin, “Image morphing using

deformable surfaces”. Proc. Computer Animation ’94, pages 31–39, 1994. IEEE Computer Society Press.

[6] S.-Y. Lee, K.-Y. Chwa, J. Hahn, and S. Y. Shin, “Image morphing using deformation techniques”. J. Visualization and Computer Animation, 7(1):3–23, 1996.

[7] S.-Y. Lee, K.-Y. Chwa, S. Y. Shin, and G. Wolberg, “Image metamorphosis using snakes and free-form deformations”. Computer Graphics (Proc.

SIGGRAPH ’95),pages 439–448,1995.

[8] J. Serra, “Image Analysis and Mathematical Morphology” vol.1, Academic Press, 1982.

[9] J. Serra, “Image Analysis and Mathematical Morphology” vol.2, Academic Press, 1988.

[10] B. Luo and E. R. Hancock, “Slice interpolation using the distance transform and morphing”. Digital Signal Processing Proceedings, 1997 13th International

Conference on Publication Date: 2-4 Jul 1997 Volume: 2, page:1083-1086 [11] R.-C. Gonzalez and R.-E. Woods, “Digital image processing second edition”

[12] G. Borgefors, “Distance transformations in digital images”. Computer Vision, Graphics, and Image Processing Pages: Volume 34 ,Issue 3 (June 1986)344 - 371 [13] G. Boreefors, “Distance Transformations in Arbitrary Dimensions”. Computer

Vision, Graphics, and Image Processing Vol. 27, NO. 3, Sept. 1984, pp. 321.345 [14] J. Serra, “Intepolations et distance de Hausdorff”, Tech.Rep. N-15/94/MM Centre

de Morpholgie Mathematique, Ecole des Mines de Paris.

[15] J. Serra, “Hausdorff distance and interpolations”. H. Heijmans and J. Roerdink in Mathematical morphology and its application to image and signal processing, Kluwer, 1998.

[16] F. Meyer, “A morphological interpolation method for Mosaic Image”, Frence [17] S. Brucher, “Interpolation of sets, of partitions and of functions”. H. Heijmans

and J. Roerdink in Mathematical morphology and its application to image and signal processing, Editors, Kluwer, 1998.

[18] S. Brucher, “Interpolation d’ensembles, de partitions ec de functions”. Tech Rep N-18/94/MM, Centre de Morphologie Mathematique, Ecole des Mines de

paris,1994

[19] Marcin Iwanowski and Jean Serra, “Morphological Interpolation and Color Images”. Proceedings of the 10th International Conference on Image Analysis and Processing, Page: 50, 1999

[20] S. P. Raya and J.K. Udupa, “Shape-based interpolation of multidimensional objects,” IEEE Trans. on Medical Imagining, vol. 1, no. 9, pp. 32–42, 1990.

[21] Gabor T. Herman, Jingsheng Zheng, and Carolyn A. Bucholtz, “Shape-based Interpolation”. IEEE Computer Graphics & Applications, 1992

using morphological morphing”. Proc. IEEE Int. Conf. Image Processing, Vol. 2, Page 1661-164, 2001

[23] Adrian G. Bors, Lefteris Kechagias, and Ioannis, “Binary Morphological Shape-Based Interpolation Applied to 3-D Tooth Reconstruction”. IEEE

TRANSACTIONS ON MEDICAL IMAGING, VOL. 21, NO. 2, FEBRUARY 2002 [24] Stanley Osher and James A. Sethian, “Fronts propagating with

curvaturedependent speed: algorithms based on Hamilton-Jacobi formulations”. J.

Comput. Phys., 79(1):12–49, 1988.

[25] Richard Tsai and Stanley Osher, “Level Set Methods in Image Science”. IEEE Image Processing, 2003. Proceedings. 2003 International Conference on Volume 2, 14-17

[26] Stanley Osher and Ronald Fedkiw, “Level Set Methods and Dynamic Implicit Surfaces”, Springer; 1 edition, New York, November 1, 2002

[27] Paul Burchard, Li-Tien Cheng, Barry Merriman, and Stanley Osher. “Motion of curves in three spatial dimensions using a level set approach”. J. Comput. Phys., 170:720–741, 2001.

[28] Li-Tien Cheng, Paul Burchard, Barry Merriman, and Stanley Osher. “Motion of curves constrained on surfaces using a level set approach”. J. Comput. Phys., 175:604–644, 2002.

[29] Baris Sumengen, “A Guide to Implementing Level Set Methods for Curve Evolution”. October 8, 2004

[30] J. A. Sethian, “Level set methods and fast marching methods: evolving interfaces in computational geometry, fluid mechanics, computer vision, and materials science”, Cambridge University Press, 1999.

[31] Stanley Osher and Nikos Paragios, “Geometric Level Set Methods in imaging, vision, and Graphics”, Springer; 1 edition , July 17, 2003

[32] Ross T. Whitaker, “A Level-Set Approach to Image Blending”. IEEE

TRANSACTIONS ON IMAGE PROCESSING, VOL. 9, NO. 11, NOVEMBER 2000 [33] David E. Breen, and Ross T. Whitaker, “A Level-Set Approach for the

Metamorphosis of Solid Models”. IEEE TRANSACTIONS ON VISUALIZATION AND COMPUTER GRAPHICS, VOL. 7, NO. 2, APRIL-JUNE 2001

[34] TURK, G., AND O’BRIEN, J. F. “Modelling with implicit surfaces that interpolate”e. ACM TOG 21(4), page 855–873. 2002.

[35] CARR, J. C., BEATSON, R. K., CHERRIE, J. B., MITCHELL, T. J., FRIGHT, W. R., MCCALLUM, B. C., AND EVANS, T. R. “Reconstruction and

representation of 3D objects with radial basis functions”. In Proc. SIGGRAPH 2001, 67–76. 2001.

[36] TURK, G., AND O’BRIEN, J. F. 1999. “Shape transformation using variational implicit functions”. In Proc. SIGGRAPH 99, 335–342.

[37] Michael Mullan, Ross Whitaker, and John C. Hart, “Procedural Level Sets”.

CARGO Remix - May 2004