Digital Signal Processing Reference
In-Depth Information
Wireless Video Measurement Based on Its Gabor
Samples in High-Dimension Space
Jun Yang
School of Computer and Information Engineering, Shanghai University of Electric Power,
Shanghai 200090, China
mail_yangjun@hotmail.com
Abstract.
Video measurement is an important issue in the wireless applications.
Video's quality is poor when it is translated by channels. Since a frame of video
can be considered as an image, its decomposed sub-images can be considered as
point set in high-dimension space. Gabor filters can be used for image mea-
surement for its biology characters. In this paper, firstly, an original image is
transformed into frequency space with different scale and angle by performing
analysis of Hue and Lightness in HSV space and combing these features in Ga-
bor space. The synthesis image is generated with different scale and angle in
different Gabor sub-space. With different sub-space, an algorithm is proposed
to calculate its measurement based on Gabor features. The proposed method is
constructive and proves the wireless video application system higher quality.
Experimental results demonstrate advantages of the proposed method over FFT
approaches.
Keywords:
video measurement, Gabor filter, high-dimension space.
1
Introduction
Wireless communication technology gives new opportunities for video applications.
Some mobile devices, such as digital television, digital cinema, wireless handset TV,
have ability to play video. Sometimes, image quality plays a key role. Due to its bio-
logical similarity to human vision system, Gabor wavelets have been widely used in
object recognition applications like fingerprint recognition [1], character recognition
[2], etc.. Despite the success of Gabor-wavelet-based object recognition systems, both
the feature extraction process and the huge dimension of Gabor features extracted
demand large computation and memory costs, which makes them impractical for real
applications[3]. In this paper, Gabor features are used to present original image. Ga-
bor wavelets seem to be the optimal basis to extract local features for image quality
presentation for several reasons:
1) Biological motivation: the shapes of Gabor wavelets are similar to the receptive
fields of simple cells in the primary visual cortex.
2) Mathematical motivation: the Gabor filters are optimal for measuring local spa-
tial frequencies.
