https://eprint.iacr.org/2023/1863
This project is the implementation of the paper "Efficient Secure Multiparty Computation for Multidimensional Arithmetics and Its Application in Privacy-Preserving Biometric Identification". The functionalities of this project include generating tensor(vector) triples, using tensor triples to compute scalar products, tensor(outer) products, and matrix multiplications. There are also examples of applications in biometric identification, including the FingerCodes and EigenFace.
Examples and benchmarks are located in src/test.cpp. The noisy functions stands for the Correlated OT(COT) based protocol in the paper, and the silent functions stands for the Silent OT(SOT). The words noisy and silent are originated from the two types of VOLE in libOTe, and we interchangeably use vt, vector triple, and tensor triple in the project and the paper.
This project depends on boost and libOTe. The author used gcc 12.2.0.
cd libOTe
python build.py --all --boost --sodium
cd ..
mkdir build
cd build
cmake ..
make -jContact [email protected] if you have any question.