Skip to content

Comparison of different face detection algoritms (in python) tested on realworld footage, but a the face was printed on paper.

Notifications You must be signed in to change notification settings

DrejcPesjak/face-detection_study

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

2 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Face detection study

Comparison of different face detection algoritms (in python) tested on realworld footage, but a the face was printed on paper.

Video files are available on my drive.

Face detectors

Detectors used: Haar cascade, Hog detector (from dlib library), DNN (from OpenCV library).

Conditions

The detectors were tested with 3 different lightning conditions: Natural daily light, Dimmed daily light in the evening and Artificial industrial light.

The tests were conducted under 5 different camera angles (90°, 75°, 60°, 45°, 30°), where 90° means straight on.

Measures

For the evaluation process the F1 score was used, which is defined as following: F1 = 2*TP / (2*TP+FP+FN)


Detailed info can be found in the report.

About

Comparison of different face detection algoritms (in python) tested on realworld footage, but a the face was printed on paper.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages