Title: Directional Gaussian filter with variable filtering strength and variable window size for deblocking digital images
Authors: Shanty Chacko; J. Jayakumar
Addresses: Department of Electronics and Communication Engineering, Karunya University, Karunyanagar, Coimbatore, 641 114, Tamil Nadu, India ' Department of Electronics and Communication Engineering, Karunya University, Karunyanagar, Coimbatore, 641 114, Tamil Nadu, India
Abstract: When a digital image is compressed at high compression ratio, visual quality of the image is affected because of the formation of compression artefacts. Blocking artefact and ringing artefact are the two major compression artefacts. In this paper, a new filtering technique is proposed to remove the blocking and the ringing artefacts from the compressed digital images. In the proposed method, a Gaussian type filter is used with variable filtering strength based on the information content in the image. For the filtering, window size of (1 × N) is used where N is a variable depending on the activities in the region. The filter is applied horizontally, vertically, along 45L and along 135L. The quality of the artefacts removed images are compared with other similar techniques. The results show that the proposed method outperforms other methods.
Keywords: blocking artefact; ringing artefact; block DCT transform; discrete cosine transform; Gaussian filters; image compression; filtering strength; window size; image deblocking; digital images; image quality.
International Journal of Information and Communication Technology, 2015 Vol.7 No.4/5, pp.425 - 436
Received: 17 Jul 2013
Accepted: 18 Sep 2013
Published online: 22 Mar 2015 *