biometricsecurity / preimage-attack-on-btp-template

Please raise your questions in the Issues section and I will take a look at the issues, do not send me emails.

Usage based on BioHashing example

Fitness function

fitness_biohashing.m

function [distance] = fitness_biohashing(x, hashcode,opts)


[transformed_data] = biohashing(x,opts.model);

% [distance] = bloomfilter_hamming(transformed_data0,hashcode,opts); %
% original distance
distcc=[];
for a=1:size(hashcode,1)
    distcc=[distcc 1 - matching_IoM(hashcode(a,:),transformed_data)];
end

distance=mean(distcc);


end

In the above code, the hashcode can be multiple templates, the fitness function will compute the fitness based on averaging

GA reconstruct set up

reconstruct.m

function [x] = reconstruct(f_fitness,f_constr,opts)
%RECONSTRUCT Summary of this function goes here
%   Detailed explanation goes here
%rng default % For reproducibility
%f_fitness = @(x)fitness(x,protected_template,opts); % fitness function
%f_constr = @(x)constraintsofx(x,protected_template,opts); % constrain function

A = [];
b = [];
Aeq = [];
beq = [];
lb = ones(1,opts.dX)*-1;
ub = ones(1,opts.dX)*1; % upper bound

%    %'PlotFcn', @gaplotbestf,...  'PlotFcn', {@gaplotbestf, @gaplotscores},

options = optimoptions('ga','ConstraintTolerance',1e-6,'FunctionTolerance',1e-10,...
    'MaxGenerations',200,'MaxTime',Inf,'CrossoverFraction',0.9,'UseParallel',true);
tic
[x,fval,exitflag,output] = ga(f_fitness,opts.dX,A,b,Aeq,beq,lb,ub,f_constr,options);
toc

end

Launch the attack

SA_BioHashing_multiple.m

 for jjj=1:5
        %% reconstruct the first one
        for i=1:158
            
            disp(['reconstructing ',num2str(i)])
            to_retrieve_hash=transformed_data((i-1)*10+1:1:(i-1)*10+jjj,:); % first of the template are used to reconstruct
            %rng default % For reproducibility
            f_fitness = @(x)fitness_biohashing(x,to_retrieve_hash,opts); % fitness function
            f_constr = []; % constrain function
            reconstruct_x(i,:) = reconstruct(f_fitness,f_constr,opts);
            
        end
         save(['data/biohashing/20190620biohashing_reconstruct_',num2str(hamming_dimension),'_',num2str(jjj),'.mat'],'reconstruct_x');
         save(['data/biohashing/20190620biohashing_eer_',num2str(hamming_dimension),'_',num2str(jjj),'.mat'],'EER_HASH');
    end

jjj indicates how many templates are compromised and are used to reconstruct the feature vector.

By downloading and using this source code, you AGREE to the terms and conditions below.

---

I agree:

1. to cite [dong-BTAS-2019] and [dong-reliability-2019] in any paper of mine or my collaborators that makes any use of the source code or data generated from these codes.
2. to use the codes and generated data for research purposes only.
3. not to provide the codes or generated data to third parties without the notice of copyright © (Dong, X., Jin, Z., Teoh, A. B. J., Tistarelli, M., & Wong, K) and this agreement.

@inproceedings{dong-BTAS-2019,
  title={A Genetic Algorithm Enabled Similarity-Based Attack on Cancellable Biometrics},
  author={Xingbo Dong and  Zhe Jin and Andrew Teoh Beng Jin},
  booktitle={10th IEEE International Conference on Biometrics: Theory, Applications and Systems (BTAS)},
  pages={},
  organization={IEEE}
}

@article{dong-reliability-2019,
      title={On the Reliability of Cancelable Biometrics: Revisit the Irreversibility},
      author={Dong, Xingbo and Jin, Zhe and Teoh, Andrew Beng Jin and Tistarelli, Massimo and Wong, KokSheik},
      journal={arXiv preprint arXiv:1910.07770},
      year={2019}
    }

---

Copyright © 2019, Dong, X., Jin, Z., Teoh, A. B. J., Tistarelli, M., & Wong, K. All rights reserved.