I have been working on Super Resolution from past few days, as it is my winter project which is supposed to be finished by the end of December. My guide told me to understand the topic and get some clear in sight on the topic by 9th of this month. so that he would give me more info on the problem statement which i have to solve.
From the past few days of my study, I have found out few interesting points super resolution and i would like to share it. For obvious reasons this post is going to be little techie.
By basic definition super resolution is the processes of obtaining high resolution image from a set of low resolution images depicting the same scene. To explain the same in little non tech language, it is a process of obtaining a high clarity image of a scene from a set of low clarity images of the same scene which were taken by a low quality camera in an interval of time.
So in order to get a high resolution image from low resolution image we usually do interpolation (bilinear, bi cubic etc). But the problem with doing interpolation is after interpolating an image we lose the sharpness of the image or the over all image looks blurred as compared to the actual scene. To avoid this problem Super resolution will be one of the best choices. In few cases super resolution can be considered as applying interpolation on more than one image.But there are several other cases where super resolution is completely different form interpolation.
Super resolution can also be applied on videos.Videos are nothing but a set of frames, and in a video each scene is depicted by a set of images. Hence we can use these set of images to get high resolution images.
This can be applied for upscaling a standard definition video into a high definition video. This is how the old movies which were shot with normal cameras can still be converted to HD.
From the past few days of my study, I have found out few interesting points super resolution and i would like to share it. For obvious reasons this post is going to be little techie.
By basic definition super resolution is the processes of obtaining high resolution image from a set of low resolution images depicting the same scene. To explain the same in little non tech language, it is a process of obtaining a high clarity image of a scene from a set of low clarity images of the same scene which were taken by a low quality camera in an interval of time.
So in order to get a high resolution image from low resolution image we usually do interpolation (bilinear, bi cubic etc). But the problem with doing interpolation is after interpolating an image we lose the sharpness of the image or the over all image looks blurred as compared to the actual scene. To avoid this problem Super resolution will be one of the best choices. In few cases super resolution can be considered as applying interpolation on more than one image.But there are several other cases where super resolution is completely different form interpolation.
Super resolution can also be applied on videos.Videos are nothing but a set of frames, and in a video each scene is depicted by a set of images. Hence we can use these set of images to get high resolution images.
This can be applied for upscaling a standard definition video into a high definition video. This is how the old movies which were shot with normal cameras can still be converted to HD.
Ryagz..
"Inspire, Innovate and Invent"
No comments:
Post a Comment