Enhanced Image Super-Resolution Technique Using Convolutional Neural Network
A framework for image super resolution using Convolutional Neural Network (CNN) was developed. In this paper, we focus on verifying the performance of Convolutional Neural Network compared to other methods. CNN generally outperforms other super resolution methods. The training images were collected from various categories, i.e. flowers, buildings, animals, vehicles, human and cuisine. The neural network trained with multiple categories of training images made the CNN more robust towards different test scenarios. Common image degradation, i.e. motion blur and noise, can be reduced when the CNN is provided with proper training samples.
KeywordsImage Super Resolution Convolutional Neural Network motion blur image denoising
Unable to display preview. Download preview PDF.
- 1.Asuni, N., Giachetti, A.: Accuracy Improvements and Artifacts Removal in Edge Based Image Interpolation. In: Proc. 3rd Int. Conf. Comput. Vis. Theory Appl. (VISAPP), pp. 58–65 (2008)Google Scholar
- 3.Tsai, R.Y., Huang, T.S.: Multiframe Image Restoration and Registration. JAI Press Inc., London (1984)Google Scholar
- 4.Tian, J., Ma, K.-K.: A Survey on Super-Resolution Imaging. Springer-Verlag London (2011)Google Scholar
- 5.Torrieri, D., Bakhru, K.: Neural Network Superresolution. IEEE, 1594–1598 (1997)Google Scholar
- 7.Sherrod, P.: DTREG, http://www.dtreg.com/mlfn.htm
- 11.Sudheer Babu, R., Sreenivasa Murthy, K.E.: A Survey on the Methods of Super-Resolution Image Reconstruction. International Journal of Computer Application 15(2) (February 2011)Google Scholar
- 12.Liu, H.Y., Zhang, Y.S., Ji, S.: Study in the Methods of Super-Resolution Image Reconstruction. The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences XXXVII(B2) (2008)Google Scholar
- 14.ICBI Page, http://www.andreagiachetti.it/icbi
- 15.The SSIM Index for Image Quality Assessment, https://ece.uwaterloo.ca/~z70wang/research/ssim/