TAILIEUCHUNG - Constructing texture maps using enhanced Beltrami method

Image quality enhancement is a crucial requirement in many applications of digital image and video processing. Removing artifacts which are suffered from image compression will lose simultaneously image texture components. | TAÏP CHÍ PHAÙT TRIEÅN KH&CN, TAÄP 19, SOÁ K2- 2016 Constructing texture maps using enhanced Beltrami method Thai Van Nguyen Tuan Do-Hong Dung Trung Vo Ho Chi Minh city University of Technology, VNU-HCM (Manuscript Received on June 16th, 2015, Manuscript Revised January 15th, 2016) ABSTRACT Image quality enhancement is a crucial requirement in many applications of digital image and video processing. Removing artifacts which are suffered from image compression will lose simultaneously image texture components. This paper combines Beltrami method and the window derivative to construct the texture map in an attempt to preserve image details during filtering artifacts. Texture map enhancement is also proposed. Simulation results show that the texture map is robust to noise and matches to real texture components of image. Key words: Standard deviation (STD), Sobel, Beltrami, texture map, window derivative. 1. INTRODUCTION Image compression is an inevitable requirement to reduce storage space of mobile devices and channel bandwidth. But compression also reduces quality of the original images. Removing artifacts and still preserving image texture is thus very important. The texture map plays an essential role in order to control the filter’s strength. The edge map guided post filters are proposed to enhance image quality in [3], [4], [6], [7]. In these methods, the variance and standard deviation operators are used to construct the edge map [10], [11], [12]. But these operators are sensitive to noise. The authors in [5] use the Sobel operator to classify edge pixels and non– edge pixels. Filtering the artifacts using this classification may blur the image due to leak of texture information. Obviously, constructing the texture map is a challenging problem since it is very difficult to define texture in mathematical terms. In [1], [2], texture feature based on the Beltrami method is used to locate texture in image segmentation. This paper constructs an enhanced .

• Kiến trúc - Xây dựng
• Constructing texture maps
• Enhanced Beltrami method
• Digital image and video processing
• Texture components of image
• Construct the texture map