Title :
Textural Adaptive Learning-Based Super Resolution for Human Face Images
Author :
Cao Yang ; Li Xiaoguang ; Li, Zhuo ; Shen Lansun
Author_Institution :
Signal & Inf. Process. Lab., Beijing Univ. of Technol., Beijing, China
Abstract :
A low-resolution face image is segmented into detailed regions and flat regions. The detailed regions are super resolved using classified predictors according to the local textural structures, while, the flat regions are magnified using bilinear interpolation. Experimental results show that both the visual quality and the computational cost are improved.
Keywords :
image resolution; image segmentation; image texture; interpolation; learning (artificial intelligence); bilinear interpolation; human face images; image segmentation; local textural structures; super resolution; textural adaptive learning; visual quality; Computational efficiency; Eyes; Face detection; Humans; Image reconstruction; Image resolution; Interpolation; Mouth; Nose; Signal resolution;
Conference_Titel :
Image and Signal Processing, 2009. CISP '09. 2nd International Congress on
Conference_Location :
Tianjin
Print_ISBN :
978-1-4244-4129-7
Electronic_ISBN :
978-1-4244-4131-0
DOI :
10.1109/CISP.2009.5304257
