Academia.eduAcademia.edu

Figure 2 – uploaded by Amisha Shah

See full PDFdownloadDownload figure
The basic premise for increasing the spatial resolution in SR techniques is the availability of multiple LR images captured from the same scene [1]. These LR images must be aliased and have sub pixel shift seeing that the LR images having shift of integer units does not enclose new information which can be used for SR reconstruction as revealed in fig.3. The SR image reconstruction consists of three stages as shown in the fig.4. By means of image registration, all pixels from available LR images can be mapped back onto the reference image. Then via interpolation a uniformly spaced up sampled image is obtained. In the end, image restoration is applied to the up-sampled image to remove the effect of sensor PSF blur and noise [4].

Figure 3 The basic premise for increasing the spatial resolution in SR techniques is the availability of multiple LR images captured from the same scene [1]. These LR images must be aliased and have sub pixel shift seeing that the LR images having shift of integer units does not enclose new information which can be used for SR reconstruction as revealed in fig.3. The SR image reconstruction consists of three stages as shown in the fig.4. By means of image registration, all pixels from available LR images can be mapped back onto the reference image. Then via interpolation a uniformly spaced up sampled image is obtained. In the end, image restoration is applied to the up-sampled image to remove the effect of sensor PSF blur and noise [4].

Related Figures (5)

Figurel. (a) Resultant HR image captured from four different cameras (b) Multiple frames captured by a video camera. restoration is used to recover a degraded image without changing its dimension.
Figurel. (a) Resultant HR image captured from four different cameras (b) Multiple frames captured by a video camera. restoration is used to recover a degraded image without changing its dimension.
Figure 4 Block diagram of constructing a SR image. Super resolution techniques are broadly classified as techniques based on frequency domain or techniques based on
Figure 4 Block diagram of constructing a SR image. Super resolution techniques are broadly classified as techniques based on frequency domain or techniques based on
Various regularization parameters In learning based approach, the relationship between an LR image and its corresponding HR image is examined via a pair of LR and HR patches. The training data then is used to predict the higher-resolution image. In order to apply the super resolution approach in real time, the super resolving time should be reduce. Xie Qinlan et al. proposed an improved example based single-image super-resolution method in [8] .To enhance the computing performance, the author had proposed a classifying method to divide the high-frequency patches of low-resolution image into different classes. For the edge-areas of the low-resolution, the routine example-based image super resolution algorithm can be used to implement the local and
Various regularization parameters In learning based approach, the relationship between an LR image and its corresponding HR image is examined via a pair of LR and HR patches. The training data then is used to predict the higher-resolution image. In order to apply the super resolution approach in real time, the super resolving time should be reduce. Xie Qinlan et al. proposed an improved example based single-image super-resolution method in [8] .To enhance the computing performance, the author had proposed a classifying method to divide the high-frequency patches of low-resolution image into different classes. For the edge-areas of the low-resolution, the routine example-based image super resolution algorithm can be used to implement the local and
Figure 5 A pictorial example of iterative back projection method. (g“ —g) in a HR image and c is a normalizing constant 1171.
Figure 5 A pictorial example of iterative back projection method. (g“ —g) in a HR image and c is a normalizing constant 1171.
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved