` tracking the eyes using a webcam presented by: kwesi ackon kwesi ackon supervisor: mr. j. connan

13
` ` Tracking the Eyes Tracking the Eyes using using a Webcam a Webcam Presented by: Presented by: Kwesi Ackon Kwesi Ackon Supervisor: Mr. J. Connan Supervisor: Mr. J. Connan

Upload: ethan-randolf-short

Post on 17-Jan-2016

217 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: ` Tracking the Eyes using a Webcam Presented by: Kwesi Ackon Kwesi Ackon Supervisor: Mr. J. Connan

Tracking the Eyes Tracking the Eyes usingusing

a Webcama WebcamPresented by: Presented by: Kwesi AckonKwesi Ackon

Supervisor: Mr. J. ConnanSupervisor: Mr. J. Connan

Page 2: ` Tracking the Eyes using a Webcam Presented by: Kwesi Ackon Kwesi Ackon Supervisor: Mr. J. Connan

RecapRecap Track the eyesTrack the eyes

Move cursor on the screen using the eyesMove cursor on the screen using the eyes

Page 3: ` Tracking the Eyes using a Webcam Presented by: Kwesi Ackon Kwesi Ackon Supervisor: Mr. J. Connan

ChangesChanges No edge detectionNo edge detection

Move from RGB colour space to HSV Move from RGB colour space to HSV colour spacecolour space

White backgroundWhite background

Page 4: ` Tracking the Eyes using a Webcam Presented by: Kwesi Ackon Kwesi Ackon Supervisor: Mr. J. Connan

HSV (Hue, Saturation, Value) SpaceHSV (Hue, Saturation, Value) Space

--H (hue): Ranges from 0-360. Hue represents theH (hue): Ranges from 0-360. Hue represents the

colour type (e.g. blue, red etc).colour type (e.g. blue, red etc).

--S (saturation): Ranges from 0-100. It represents the purity of the S (saturation): Ranges from 0-100. It represents the purity of the colour: the higher saturation value is, the clearer the colour is. If colour: the higher saturation value is, the clearer the colour is. If saturation is low, the colour looks closer to gray.saturation is low, the colour looks closer to gray.

--V (value): This is the brightness of the colour, and it ranges from V (value): This is the brightness of the colour, and it ranges from 0-100.0-100.

Page 5: ` Tracking the Eyes using a Webcam Presented by: Kwesi Ackon Kwesi Ackon Supervisor: Mr. J. Connan

HSV colour space represented as a cylinderHSV colour space represented as a cylinder

Red pixelRGB space: 255 0 0HSV space: 0˚ 100% 100%

Page 6: ` Tracking the Eyes using a Webcam Presented by: Kwesi Ackon Kwesi Ackon Supervisor: Mr. J. Connan

Conversion from RGB to HSVConversion from RGB to HSV

Page 7: ` Tracking the Eyes using a Webcam Presented by: Kwesi Ackon Kwesi Ackon Supervisor: Mr. J. Connan

ProgressProgress Face detection implementationFace detection implementation

Original Original image Detected skin pixelsDetected skin pixels0 < H < 50, 23 < S < 68 0 < H < 50, 23 < S < 68

Level of accuracy 96.4%

Page 8: ` Tracking the Eyes using a Webcam Presented by: Kwesi Ackon Kwesi Ackon Supervisor: Mr. J. Connan

Results for different skin typesResults for different skin types

Dark skinDark skin

Light skinLight skin

Page 9: ` Tracking the Eyes using a Webcam Presented by: Kwesi Ackon Kwesi Ackon Supervisor: Mr. J. Connan

Eye detection implementationEye detection implementation

yx

y1

x1

Obtaining dimensionsof eye rectangle:y = Half height of face rectanglex = Half width of face rectangley1 = 0.10 * yx1 = 0.625 * x Width of eye rectangle = x1Height of eye rectangle= 2 * y1

Page 10: ` Tracking the Eyes using a Webcam Presented by: Kwesi Ackon Kwesi Ackon Supervisor: Mr. J. Connan

Results for eye detectionResults for eye detection

Page 11: ` Tracking the Eyes using a Webcam Presented by: Kwesi Ackon Kwesi Ackon Supervisor: Mr. J. Connan

Tools and LanguagesTools and Languages Dev-C++ 4.9.9.2Dev-C++ 4.9.9.2 Visual C++ 6.0Visual C++ 6.0 OpenCV 1.0.0 OpenCV 1.0.0

Page 12: ` Tracking the Eyes using a Webcam Presented by: Kwesi Ackon Kwesi Ackon Supervisor: Mr. J. Connan

PlanPlan Move cursor on screen using the eyesMove cursor on screen using the eyes

Page 13: ` Tracking the Eyes using a Webcam Presented by: Kwesi Ackon Kwesi Ackon Supervisor: Mr. J. Connan