14-11-2012, 01:56 PM
DEMOSAICKING BASED ON IMAGE REFINEMENT USING GRADIENTS AND ARTIFACT REDUCTION USING WAVELETS
demosaicking_balanivetha.pdf (Size: 123.85 KB / Downloads: 32)
Introduction:
Recent advances in hardware, digital image processing and telecommunications have allowed for the
miniaturization and incorporation of digital cameras in consumer electronic devices, such as mobile phones and
wireless PDA’s. A color image usually consists of three channels per pixel, each carrying the information of a
specific wavelength sensitivity band (red, green, or blue) to allow for color processing and display. Intuitively,
one would assume that acquiring such an image requires cameras with three spatially aligned sensors, each
preceded by a different color filter to capture the information for a given part of the visible spectrum. However,
to reduce size, cost, and image registration errors, most digital cameras only have a single sensor with a color
filter array (CFA) placed in front of it. The specific arrangements of color filters in the CFA vary between the
camera manufacturers which use not only RGB CFAs, however, the patterns with complementary Cyan ©,
Magenta (M), Yellow (Y) colors, or four-color CFAs formed through mixed primary (RGB) and complementary
(CMY) colors are in the use today as well. Among these, the Bayer pattern (Fig.1) [1] is commonly used due to
simplicity of the subsequent processing steps.
Demosaicking Methods:
An immense number of demosaicking techniques have been proposed to address CFA demosaicking over the
years. The simplest one is probably bilinear interpolation [2], which fills missing color values with weighted
averages of their neighboring pixel values. Although computationally efficient and easy to implement, it
introduces several demosaicking artifacts and smears sharp edges. To obtain more visually pleasing results,
many adaptive CFA demosaicking methods have been proposed. In general, they utilize the information of intra
and inter channel correlations. The intra channel correlation techniques use the continuity of pixel intensity in
each color channel to perform interpolation. These techniques include gradient based interpolation [8] and other
correlation based interpolation methods,[13],[14],[17]. Furthermore, it is well known that there exists a certain
extent of correlation between green, red and blue channels.
Experimental Results:
This section is devoted to the detailed performance comparison. The algorithm is simulated for a list of Kodak
test images shown in Fig.4. These images are film captures and digitized with photo scanners. They are chosen
for the experiments as the images contain a significant amount of color texture, which can be used to determine
the effectiveness of various demosaicking algorithms. The CFA images are simulated by sampling the original
images using the Bayer pattern. The most commonly used metric for evaluating the quality of the proposed
algorithm are Mean square Error (MSE) and Peak Signal to Noise Ratio (PSNR).The results of the proposed
method is compared with the results of five benchmark methods namely the Bilinear Interpolation, Alternating
Projections [6],Optimal Recovery [16],Successive Approximation [10] and Difference Plane Model [15].