Research Article
Arabic Sign Language Recognition System for Alphabets Using Machine Learning Techniques
Table 7
The hand features used in AArSLRS.
| Features | Feature description | Feature equation |
| F1 | The ratio coordinates of the hand center of mass on the x-axis, to the width of around box by hand | F1 = CX/x.width |
| F2 | The ratio coordinates of the hand center of mass on the y-axis, to the length of around box by hand | F2 = CY/y.length |
| F3 | The ratio between the width of around box by hand to the length of around box by hand | F3 = x.width/y.length |
| F4 | The ratio pixels of between the object (hand) area to the area of around box by hand | F4 = area (object)/area of bounding box |
| F5 | The ratio between object (hand) perimeter to object (hand) area | F5 = perimeter (object)/area (object) |
| F6 | The ratio between object major axis length to area of bounding box | F6 = object major axis length/area of bounding box |
| F7 | The ratio between object minor axis length to area of bounding box | F7 = object minor axis length/area of bounding box |
| F8–F15 | Extrema (endpoints of hand shape) | F8 = X top-left | F9 = X top-right | F10 = X right-top | F11 = Y right-top | F12 = X right-bottom | F13 = Y right-bottom | F14 = Y left-bottom | F15 = Y left-top |
|
|