HandGesturePy

Human Computer Interaction using hand gesutures using Opencv - Python . Hand Gesture recognition is performed using HoG features and SVM as classifier. This project is inspired from http://research.microsoft.com/pubs/220845/depth4free_SIGGRAPH.pdf

Hardware Setup

Difference after removing IR illuminator and using IR illuminators

Final Setup

Project Modules

TrainData

Training Data Accuracy

Number of Class : 9
Number of Item per Class : 20
Folder Name : TrainData2
Number of Cross validation Folds : 20

========================================================
+													   +
+				    MODEL EVALUATION				   +
+													   +
========================================================


----------------------- accuracy -----------------------
Mean    :    0.938888888889
Max     :    1.0
Min     :    0.666666666667
----------------------- f1_macro -----------------------
Mean    :    0.921296296296
Max     :    1.0
Min     :    0.611111111111
----------------------- f1_micro -----------------------
Mean    :    0.938888888889
Max     :    1.0
Min     :    0.666666666667
----------------------- precision_macro ----------------
Mean    :    0.912962962963
Max     :    1.0
Min     :    0.592592592593
----------------------- precision_micro ----------------
Mean    :    0.938888888889
Max     :    1.0
Min     :    0.666666666667
----------------------- precision_weighted -------------
Mean    :    0.912962962963
Max     :    1.0
Min     :    0.592592592593
----------------------- recall_macro -------------------
Mean    :    0.938888888889
Max     :    1.0
Min     :    0.666666666667
----------------------- recall_micro -------------------
Mean    :    0.938888888889
Max     :    1.0
Min     :    0.666666666667
----------------------- recall_weighted ----------------
Mean    :    0.938888888889
Max     :    1.0
Min     :    0.666666666667