BEAT - robertodaza/kboc16_baseline

This library does not use any other libraries.

import scipy as sc

import numpy as np

from scipy import linalg

import bob
    
def classifier_knn_norm(gallery_sample,test_sample):#This is the Combined Manhattan-Mahalanobis distance
    vec_cov=np.linalg.pinv(np.cov(gallery_sample,rowvar=0))#  pinv returns the pseudoinverse
    vec_cov=sc.real(linalg.sqrtm(vec_cov+0.001))
    final=np.size(gallery_sample,axis=0)
    for i in range(0,final):
        if i==0:
            gallery_normalized=sc.transpose(np.dot(vec_cov,sc.transpose(gallery_sample[i])))
        else:
            gallery_normalized=np.vstack([gallery_normalized,sc.transpose(np.dot(vec_cov,sc.transpose(gallery_sample[i])))])
    f1=sc.transpose(np.dot(vec_cov,sc.transpose(test_sample))) 
    d2=np.zeros((final,1))
    try:
        for i in range(0,final):
            d2[i]=sc.sum(abs(f1-gallery_normalized[i]))
        score=np.min(d2)+0.001
    except:
        score=-1
    return score    
 
def classifier_knn_mahal(gallery_sample,test_sample):#This is the Mahalanobis + Nearest Neighbor distance(Classification of keystroke timing based on knn and cityblock distance)
    gallery_sample=np.asarray(np.asmatrix(gallery_sample))
    test_sample=np.asarray(np.asmatrix(test_sample))
    try:
        vec_cov=np.linalg.pinv(np.cov(gallery_sample,rowvar=0))
        vec_cov=sc.real(linalg.sqrtm(vec_cov+0.001))
        #vec_cov=sc.real(vec_cov+0.001)
        final=np.size(gallery_sample,axis=0)
        d3=np.zeros((final,1))
        for i in range(0,final):
            d3[i]=np.dot(np.dot(test_sample-gallery_sample[i],vec_cov),sc.transpose(test_sample-gallery_sample[i]))
        score=np.min(d3)+0.001    
    except:
        score=-1
    return score

def classifier_manhattand_scaled_modified(gallery_sample,test_sample):# This is Modified Scaled Manhattan distance and a normalization  by standard deviation
    gallery_sample=np.asarray(np.asmatrix(gallery_sample))
    test_sample=np.asarray(np.asmatrix(test_sample))
    medias_train = sc.mean(gallery_sample,axis=0)
    sigmas_train = sc.std(gallery_sample,axis=0) 
    umbral_sigmas = 0.2*np.mean(sigmas_train)
    sigmas_train[np.nonzero(sigmas_train<umbral_sigmas)]=umbral_sigmas;
    if np.size(gallery_sample,axis=1)==np.size(test_sample,axis=1):
        score=np.mean(abs(test_sample-medias_train)/sigmas_train)
    else:
        score=-1
    return score

The code for this algorithm in Python
The ruler at 80 columns indicate suggested POSIX line breaks (for readability).
The editor will automatically enlarge to accomodate the entirety of your input
Use keyboard shortcuts for search/replace and faster editing. For example, use Ctrl-F (PC) or Cmd-F (Mac) to search through this box

Three baseline systems (Combined Manhattan-Mahalanobis distance,Mahalanobis + Nearest Neighbor,Modified Scaled Manhattan distance)

Algorithms 6

Privacy:

Updated	Name	Actions
Aug. 21, 2018	loginofdeath/kboc_neoeyed/1	1
May 5, 2016	Alberto/proof_1/1 (KBOC16: Participant block)	1
April 7, 2016	anjithgeorge/KBOC_IITKGP_ANJITH_APRIL7_1/14 (KBOC16: Participant block)	2
March 8, 2016	robertodaza/KBOC16_PARTICIPANT_BLOCK/2 (KBOC16: Participant block)	0
March 2, 2016	aythamimm/KBOC16_UAM_Tutorial_Example/1 (KBOC16: Participant block)	1
Feb. 13, 2016	robertodaza/KBOC16_PARTICIPANT_BLOCK/1	2

This table shows the number of times this library has been successfuly used at a given environment. Note this does not provide sufficient information to evaluate if the library will work when submitted to different conditions.

libraries robertodaza kboc16_baseline_matchers 5

Algorithms 6

libraries

robertodaza

kboc16_baseline_matchers

5