840 C.Liu et al. d 9 h Fig.7a-h.Profile of a teaching building.a Target image.b LDV image.c,d Warped PSRs from the LDV image.e Optimal seam lines shown as the contours of the green and yellow regions.f Repairing result without Poisson image blending.g The final repairing result. h The original scene 5 Conclusion and future work user interaction,which is time-consuming.We wish to employ the information of color distribution and rela- Image completion based on views of large displacement tive location of the initially segmented image patches fills in occluded or damaged regions of a target image ex- to automatize this process. ploiting the visible information of the LDV images.In this At present,we only verify our method with a single paper,we propose an effective solution for this problem. LDV image.However,image completion from multi- By employing the techniques of image segmentation and ple images with different views will be more useful image matching,we transform the LDV image into us- for photo editing and completion.We shall further test able information,and then repair the missing region with the cases of several LDV images with complex occlu- a new image repairing algorithm.Experiments show that sions. our method works well even for large missing regions with complex structure,and achieves superior repairing results to previous image completion techniques Acknowledgement This work was supported by the Natural Sci- The following topics are to be investigated in future: ence Foundation of China (grant nos.60403038 and 60603076) and the National Basic Research Project of China (grant no. PSRs on the target image and the LDV image one as 2002CB312101).The authors are grateful for the stimulating dis- well as their counterparts are currently specified by cussions with Drs.Jin Wang,Wei Chen and Changbo Wang. References 1.Ballester,C..Caselles,V..Verdera.J.. 4.Boykov,Y..Kolmogorov.V.:An 8.Criminisi,A..Perez.P..Toyama,K.: Bertalmio.M..Sapiro.G.:A variational experimental comparison of min-cut/max- Region filling and object removal by model for filling-in gray level and color flow algorithms for energy minimization in exemplar-based image inpainting.IEEE images.In:Proceedings of the IEEE ICCV vision.IEEE Trans.Patter Anal.Mach. Trans.Image Processing 13(9),1200-1212 2001.vol.1,Pp.10-16(2001) ntell.269).1124-1137(2004) (2004) 2.Bertalmio,M..Bertozzi A.L.,Sapiro,G.: 5.Chan,T.,Shen,J.H.:Variational image 9.Drori,I.,Cohen-Or,D..Yeshurum,H.: Navier-Stokes,fluid dynamics,and image inpainting.Commun.Pure Appl.Math. Fragment-based image completion.ACM and video inpainting.In:Proceedings of the 58(5).579-619(2005) Trans.Graph.22(3).303-312 IEEE CVPR 2001,vol.I,pp.355-362 6.Collis,B..Kokaram.A.:Filling in the (2003) (2001) gaps.IEE Electron.Systems Software 2(4). 10.Fadili,M.J.,Starck,J.L.:EM algorithm for 3.Bertalmio,M.,Sapiro.G.,Caselles,V.. 22-28(2004) sparse representation-based image Ballester,C.:Image inpainting.In: 7.Comaniciu,D.,Meer,P.:Mean Shift: inpainting.In:Proceedings of the IEEE Proceedings of ACM SIGGRAPH 2000 A robust approach toward feature space ICP2005,vol.2,pp.61-64(2005) pp.417-424.New Orleans,Louisiana analysis.IEEE Trans.Pattem Anal.Mach. 11.Hartley,R.,Zisserman,A.:Multiple view (2000) Intell.245),603-619(2002) geometry in computer vision.Cambridge840 C. Liu et al. Fig. 7a–h. Profile of a teaching building. a Target image. b LDV image. c,d Warped PSRs from the LDV image. e Optimal seam lines shown as the contours of the green and yellow regions. f Repairing result without Poisson image blending. g The final repairing result. h The original scene 5 Conclusion and future work Image completion based on views of large displacement fills in occluded or damaged regions of a target image ex￾ploiting the visible information of the LDV images. In this paper, we propose an effective solution for this problem. By employing the techniques of image segmentation and image matching, we transform the LDV image into us￾able information, and then repair the missing region with a new image repairing algorithm. Experiments show that our method works well even for large missing regions with complex structure, and achieves superior repairing results to previous image completion techniques. The following topics are to be investigated in future: – PSRs on the target image and the LDV image one as well as their counterparts are currently specified by user interaction, which is time-consuming. We wish to employ the information of color distribution and rela￾tive location of the initially segmented image patches to automatize this process. – At present, we only verify our method with a single LDV image. However, image completion from multi￾ple images with different views will be more useful for photo editing and completion. We shall further test the cases of several LDV images with complex occlu￾sions. Acknowledgement This work was supported by the Natural Sci￾ence Foundation of China (grant nos. 60403038 and 60603076) and the National Basic Research Project of China (grant no. 2002CB312101). The authors are grateful for the stimulating dis￾cussions with Drs. Jin Wang, Wei Chen and Changbo Wang. References 1. Ballester, C., Caselles, V., Verdera, J., Bertalmio, M., Sapiro, G.: A variational model for filling-in gray level and color images. In: Proceedings of the IEEE ICCV 2001. vol. 1, pp. 10–16 (2001) 2. Bertalmio, M., Bertozzi A.L., Sapiro, G.: Navier–Stokes, fluid dynamics, and image and video inpainting. In: Proceedings of the IEEE CVPR 2001, vol. I, pp. 355–362 (2001) 3. Bertalmio, M., Sapiro, G., Caselles, V., Ballester, C.: Image inpainting. In: Proceedings of ACM SIGGRAPH 2000, pp. 417–424. New Orleans, Louisiana (2000) 4. Boykov, Y., Kolmogorov, V.: An experimental comparison of min-cut/max- flow algorithms for energy minimization in vision. IEEE Trans. Pattern Anal. Mach. Intell. 26(9), 1124–1137 (2004) 5. Chan, T., Shen, J.H.: Variational image inpainting. Commun. Pure Appl. Math. 58(5), 579–619 (2005) 6. Collis, B., Kokaram, A.: Filling in the gaps. IEE Electron. Systems Software 2(4), 22–28 (2004) 7. Comaniciu, D., Meer, P.: Mean Shift: A robust approach toward feature space analysis. IEEE Trans. Pattern Anal. Mach. Intell. 24(5), 603–619 (2002) 8. Criminisi, A., Perez, P., Toyama, K.: ´ Region filling and object removal by exemplar-based image inpainting. IEEE Trans. Image Processing 13(9), 1200–1212 (2004) 9. Drori, I., Cohen-Or, D., Yeshurum, H.: Fragment-based image completion. ACM Trans. Graph. 22(3), 303–312 (2003) 10. Fadili, M.J., Starck, J.L.: EM algorithm for sparse representation-based image inpainting. In: Proceedings of the IEEE ICIP 2005, vol. 2, pp. 61–64 (2005) 11. Hartley, R., Zisserman, A.: Multiple view geometry in computer vision. Cambridge