Abstract

Image denoising is a very important technique and has been widely used in many image processing applications. A low complexity image processing algorithms are necessary for VLSI implementation or real time applications, low complexity image processing algorithms are necessary for VLSI implementation. The existing high quality image denoising algorithms have characteristics of high complexity and high memory requirement which is not easily realized by VLSI technique. In order to eliminate the drawbacks of the existing algorithms we are proposing a new adaptive algorithm known as decision tree based detection method (DTBDM) for removing the random valued impulse noise in images. It is a low complexity algorithm. Here firstly, the erroneous pixels are detected by using the decision tree based algorithm and an edge preserving image filter and a simple average filter are used to reconstruct pixel value corresponding to the erroneous pixels. And these reconstructed pixels are adaptively written back as a part of the input data. The proposed design requires simple computations and two line buffers only, so hardware cost is low. Here in our design 99.6 percent of storage is reduced. The proposed system can remove the corrupted images efficiently and the performance is comparable with the high complexity techniques. The hardware requirement can be reduced further by modifying the design of the edge preserving image filter

Paper Details