Panorama stitching or image stitching is gaining prominence in today's world due to its applications in wide context aware imaging. The project aims on three different image stitching techniques and comparing their efficiencies and computational time. After the image is acquired, if the image parameters are unknown, feature detection and extraction is carried out without undistortion and cylindrical warping. In this case, the SIFT features are calculated from the acquired images and the images are translated based on the obtained homography. Then images are blended using alpha or weighted blending. If the image parameters, such as focal length and intrinsic parameters are known after image acquisition, the image first undergoes undistortion and warping. After that, two approaches are followed to compare efficiencies between the methods. The first approach involves the direct method in which the translation between the images is calculated with the need for features to reduce the calculated time manifolds and aid in faster translation approximation. And also feathering is performed to blend the images seamlessly and obtain the stitched image. In the second approach, the feature based method is used wherein the SIFT features are calculated for the image sets. Following this, the features are used to calculate the homography between the two images and determine the translation. Then, the images are stitched using a different blending technique called pyramid blending. This technique proves slightly more efficient than the previous technique due to computation of Laplacian and Gaussian pyramids of the images. © 2006-2014 Asian Research Publishing Network (ARPN).