Images are playing very key role in various applications like medical imaging, remote sensing, and scientificresearch and in all these applications image clarity plays a major role. Any image that is captured has to be processed for noise removal because noise in image is unavoidable and mixes with image either during capturing or during transmission. Various image denoising methods are developed earlier and are also working for better one. Among various methods developed Sparsity based Non local means methods are giving better denoising results. Main drawbacks of image denoising methods are they smoothens the fine details of the denoised image. This drawback can be overcome by using gradient histogram as reference while restoring the image. This paper discuss about some of gradient based image denoising methods and their comparative results.