User authentication is an indispensable part of a secure system. The traditional authentication methods have been proved to be vulnerable to different types of security attacks. Artificial intelligence is being applied to crack textual passwords and even CAPTCHAs are being dismantled within few attempts. The use of graphical password as an alternate to the textual passwords for user authentication can be an efficient strategy. However, they have been proved to be susceptible to shoulder surfing like attacks. Advanced authentication systems such as biometrics are secure but require additional infrastructure for efficient implementation. This paper proposes a novel pattern-based multi-factor authentication scheme that uses a combination of text and images resulting for identifying the legitimate users. The proposed system has been mathematically analyzed and has been found to provide much larger password space as compared to simple text based passwords. This renders the proposed system secure against brute force and other dictionary based attacks. Moreover, the use of text along with the images also mitigates the risk of shoulder surfing.