In this paper, we suggest to improve the SIRB (SIFT (Scale-Invariant Feature Transform) and ORB (Oriented FAST and Rotated BRIEF)) algorithm by incorporating RANSAC to enhance the matching performance. We use multi-scale space to extract the features which are impervious to scale, rotation, and affine variations. Then the SIFT algorithm generates feature points and passes the interest points to the ORB algorithm. The ORB algorithm generates an ORB descriptor where Hamming distance matches the feature points. We propose to use RANSAC (Random Sample Consensus) to cut down on both the inliers in the form of noise and outliers drastically, to cut down on the computational time taken by the algorithm. This postprocessing step removes redundant key points and noises. This computationally effective and accurate algorithm can also be used in handheld devices where their limited GPU acceleration is not able to compensate for the computationally expensive algorithms like SIFT and SURF. Experimental results advocate that the proposed algorithm achieves good matching, improves efficiency, and makes the feature point matching more accurate with scale in-variance taken into consideration.