<p>Image denoising is an important precursor in many computer vision problems. Any fractional integral based denoising filter needs to be optimized with respect to a set of parameters that generally include the order of integration. Unfortunately, even for many state-of-the-art fractional filters, this is done by brute experimentation over a large range of values. This process is laborious and computationally expensive. The contribution of the proposed work is twofold. Firstly, we propose an algorithm to search for an optimal order of integration in a denoising fractional filter. Secondly, deploying the same algorithm, we present a comparative analysis of seven fractional filters under the framework of image denoising. The comparison of the denoising performance of these filters is carried out for three different types of noise, namely, salt-and-pepper noise, additive Gaussian noise, and multiplicative noise. Our study of different denoising filters at varying noise levels shows that the choice of filter depends on the degree of noise variance or density. For higher variance noise, filters derived from the Reimann–Liouville fractional integral operator, K-operator, and left/forward-weighted fractional integral operator perform better than the other fractional filters.</p>