Skip to content

Training, evaluating and choosing the best model for sign recognition (signs from Polish Sign Language)

Notifications You must be signed in to change notification settings

annasli378/ChoosingModelForPSLRecognition

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 

Repository files navigation

ChoosingModelForPSLRecognition

Training, evaluating and choosing the best model for sign recognition (signs from Polish Sign Language)

Training dataset

Training dataset consists of 29 groups corresponding with choosen letters, numbers and additional signs from PSL.

Class labels:

Data re-processing consist of:

  • determining 21 hand landmarks using mediapipe library with the minimum detection confidence = 0.7
  • normalization of input data to the range (0, 1)
  • obtaining 42 coordinates (x, y)

Algorithms

Support Vector Machine

K-nearest neighbors

Random Forest Algorithm

Results

Bibliography

About

Training, evaluating and choosing the best model for sign recognition (signs from Polish Sign Language)

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published