One well-studied image processing task is the remova l of impulse noise from images. Impulse noise can be introduced during image capture, during transmission, or during storage. The signal-dependent rank order mean (SD-ROM) filter has been shown to be effective at removing impulses from 2-D scalar-valued signals while preserving many details and other features. The algorithm is based on state-estimation approach where first the state of each sample is determined and then an appropriate replacement is calculated, if required. Excellent results have been obtained for both two-state and multi- state versions of the filter. However, many images re not scalar-valued, but rather vector-valued. For example, color images have three values associated with each sample location. The scalar SD-ROM algorithm uses sorted window values to estimate the state of sample. There are many ways to sort vectors. In this paper, we describe the strengths of the scalar SD-ROM algorithm and the challenges inherent in developing vector SD-ROM algorithm. We will also present simulation results obtained using corrupted color images for proposed vector impulse-removing algorithm.