Low Rank Approximation - using SVD Low rank approximation is reducing the matrix to its nearest low rank by reducing singular values till to the rank we need to approximate even it applies to unitary vectors U and V.
Image is a matrix with each element which represents pixel intensities from 0 to 255. If its a colored image, then it contains three layers representing - RGB.
To compress image we use lower approximation of matrix and then save it which will result in compressed image.
This program returns compressed image. -compressImage('TestImage.jpg',80)