Date of Publication :21st February 2018
Abstract: Many places e.g. parks, gardens, zoos and historical places are guarded by fences. The images captured at these places are occluded with a fence. This affects the authentic appeal of the image. So, removal of the fence form those images are required. In this paper, a novel algorithm is proposed for the removal of the fence from the digital images. The fence in the image is segmented by using the multiple threshold values. The segmented fence is used as a mask for the restoration of the image from the fence occlusion. The fence occluded region is restored by using the hybrid inpainting technique. The proposed algorithm is tested on a wide verity of images. Comparative results are provided to demonstrate the effectiveness of the proposed algorithm.
Reference :
-
- T. Tamaki and H. Suzuki, “String-like occluding region extraction for background restoration,” in ICPR, vol. 3. IEEE, 2006, pp. 615–618.
- Liu, Y., Belkina, T., Hays, J., Lublinerman, R.: Image defencing. In: Proceedings of the IEEE CVPR, pp. 1–8 (2008)
- Criminisi, A., Perez, P., Toyama, K.: Region filling and object removal by exemplar-based image inpainting. IEEE Trans. Image Process. 13(9), pp. 1200–1212 (2004)
- Park, M., Brocklehurst, K., Collins, R.T., Liu, Y.: Image de-fencing revisited. In: Proceedings of the Asian Conference on Computer Vision (ACCV), pp. 422–434 (2011)
- Hettiarachchi, R., Peters, J., Bruce, N.: Fence-like quasiperiodic texture detection in images. Theory and Applications of Mathematics & Computer Science 4 (2), pp. 123-139 (2014)