The increasing case of malignant melanoma across the world compels the need for more focus on diagnosing it in early stage. A survey of the last five years shows the differences in the survival rate between the treated people of the advanced stage (15%) and the early stage (95%), and it shows the importance of the early detection of melanoma and early treatment. Non-invasive methods like dermoscopy are popular due to its accuracy but still, it is time-consuming and complex for interpretation. To overcome this, computer aided analysis methods were developed and it helped vastly to identify and analyze whether the affected lesion is melanoma or benign in the early stage. For computer Aided Diagnosis (CAD), a clear image without noise and artifacts are needed for accurate interpretation. Hence the importance of image processing plays an important role in these automated analyzing methods. This paper attempts to discuss and review about the techniques and steps involved in image processing and machine learning for skin cancer analysis. © 2019, Institute of Advanced Scientific Research, Inc.. All rights reserved.