Partial fingerprint identification systems recognise an individual when the sensor size has a small form factor in accepting a full fingerprint. However, the distinctive features within a partial fingerprint are significantly less. Hence, the uniqueness of a partial fingerprint cannot be assured, leading to the possibility of identifying multiple users. A MasterPrint is a partial fingerprint identifying at least 4% distinct individuals in a partial fingerprint identification system. This work addresses the MasterPrint vulnerability by proposing a novel partial fingerprint identification scheme that extracts minutiae-oriented local features from binarized and thinned partial fingerprint images over eight axes emerging from a reference minutia. It also introduces a metric to compute the similarity score between two partial fingerprint templates. The results are compared with the baseline minutiae matching (BMM) method, a modified Speeded-Up Robust Features (SURF) based approach, VeriFinger 12.1 SDK, standard NIST NBIS, and Ridge Shape Feature (RSF) scheme. The experiments employing partial fingerprint datasets cropped from standard FVC2002 DB1_A, FVC2002 DB2_A, NIST Special Databases (sd302b and sd302d), and CrossMatch VeriFinger dataset have demonstrated that the proposed method generates the lowest MasterPrints with the highest identification accuracy.