Image registration, a process of mapping two or more images having different alignment but sharing the same information, is important step in many applications. These applications require perceptible data from different images for comparison, integration or analysis. Image registration plays important role in medical imaging as most of the applications is related to clinical prognosis. Here, four image registration techniques namely geometric transformation, phase correlation, text{log} polar transform and adaptive polar transform are analyzed. Registration algorithms measure transformation to set correspondence between more than two images. The purpose of this proposal is to give in depth analysis of all the techniques and observe which gives best mapping results with the two images, and also find parameters which gives relevant information which helps to medical experts. Rotation and scaling factor are obtained by log polar transform. But because of non-uniform sampling, LPT is not proper approach for some applications. Therefore, APT is used which maintain the robustness to scale and rotation. For translation property, phase correlation help. However, limitations of all these techniques are still exit. © 2018 IEEE.