We present a two-party secure information processing protocol referred to as SIPPA-2.0 - targeted towards privacy preserving biometric data comparison and reconstruction. The original intention of SIPPA as reported previously is to enable private data comparison and reconstruction between a client and a server when (a) the client possesses some data that are “sufficiently similar” to that of the server, and (b) the server provides a scalar helper data to facilitate private data reconstruction by the client. In SIPPA-2.0, private data comparison and reconstruction are based on new theoretical results and a novel secure computation protocol referred to as SLSSP. These new results allow us to design and develop the much improved SIPPA and SLSSP protocols guaranteeing (a) security under semi-malicious model rather than just semi-honest model, and (b) privacy assurance with arbitrary reconstruction accuracy controllable by the server. Security analysis proving SLSSP secure under the semi-honest and semi-malicious models is presented. SIPPA-2.0 is applied to enable privacy preserving fingerprint comparison; where two parties can compare their fingerprint samples and can obtain a similarity score without revealing their raw fingerprint to each other. Experimental results on the accuracy of fingerprint matching and the run-time performance are also reported.