Fingerprint-based biometric recognition is integral to identity verification in various domains, including security, transaction, attendance, and forensic investigation. Hence, making this process efficient and robust is an essential field of research. The efficiency of any fingerprint-based biometric recognition system is highly dependent on the feature extraction process. Existing works on indirect feature extraction from fingerprints are either computationally expensive or not scalable. A novel idea for indirect feature extraction from fingerprints is discussed in this chapter. The extracted indirect features are based on the fingerprint’s minutiae points and are proved to be scale, rotation, and translation invariant. The given idea is scalable and computationally efficient while robust to noise and distortion. The extracted features have high intra-class similarity and vice versa, which is desirable for any biometric recognition system. The proposed approach is evaluated using benchmark FVC datasets and a large dataset generated synthetically. The results show improved matching accuracy and lower time complexity over the existing methods.