Advanced Discriminative Transfer Learning for General Image Restoration
S P Maniraj1, Manonmani Lakshmanan2, Siddhartha Roy3, Moumita Dutta4, Aastha Sharma5

1S.P. Maniraj, Assistant Professor, (Senior Grade), SRM Institute of Science and Technology, Chennai (Tamil Nadu), India.
2Manonmani Lakshmanan, B. Tech. UG Scholar, SRM Institute of Science and Technology, Chennai (Tamil Nadu), India.
3Siddhartha Roy, B. Tech. UG Scholar, SRM Institute of Science and Technology, Chennai (Tamil Nadu), India.
4Aastha Amar Sharma, B. Tech. UG Scholar, SRM Institute of Science and Technology, Chennai (Tamil Nadu), India.
5Moumita Dutta, B. Tech, UG Scholar, SRM Institute of Science and Technology, Chennai (Tamil Nadu), India.

Manuscript received on 18 April 2019 | Revised Manuscript received on 25 April 2019 | Manuscript published on 30 April 2019 | PP: 325-327 | Volume-8 Issue-4, April 2019 | Retrieval Number: D6110048419/19©BEIESP
Open Access | Ethics and Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)

Abstract: Low light images involve various techniques to procure a clear ground image. These images have various other disruptions such as mosaicking, blurring, etc. yet prevail in the image. To overcome this, the Advanced Discriminative Transfer Learning (ADTL) is proposed. This method uses novel approaches such as Discriminative transfer learning and by taking Synthetic Aperture Radar (SAR) images. Initially, the pre-processing of the image is done by increasing the intensity of the image. To the SAR images, the speckle reduction algorithm is applied, wavelet noise threshold is added, image de-noising and image reconstruction is done. Data proximal operator is used that helps a wide range of images fit into one common algorithm in DTL. Under DTL, various functionalities such as de-mosaicing, de-blurring, in-paint, de-noising, etc. is used to recover the disrupted image. These techniques help improve the quality of the image and produces a resultant image which is close to the ground image in a short span of time.
Keywords: Advance Discriminative Transfer Learning (ADTL), Discriminative Transfer Learning (DTL), Synthetic Aperture Radar (SAR)

Scope of the Article: Image Processing