1. Gender Classification Using Scaled Conjugate Gradient Back Propagation
Computer Science Department
Gender Classification Using Scaled Conjugate Gradient Back Propagation
Authors : Assist.Prof. Dr. Abbas H. Hassin Al-Asadi.
: Entesar Barges Talal.
Date : 25/2/2016.

2. Outlines Introduction The Proposed Algorithm Conclusion
Suggestion for future works

3. Introduction Gait:-a particular way or manner of moving on foot.
Gait cycle:  The gait cycle is the time interval between the exact same repetitive events of walking. The defined cycle can start at any moment, but it is generally begin when one foot contacts the ground. If it starts with the right foot contacting the ground, then the cycle ends when the right foot makes contact again.
  There are two phases of the gait cycle:
Stance phase.
Swing phase.
Gait recognition :The term gait recognition signifies the identification of an individual from a video sequence of the subject walking. This does not mean that gait is limited to walking, it can also be applied to running or any means of movement on foot.

4. Gender Classification
  Gender classification :is to determine a person's gender, e.g., male or female, based on his or her biometric cues. In this correspondence, we present method of gender classification based on human gait.

5. The Proposed Algorithm
Input Video
Frames
Eigenbackgrounds
Preprocessing
If mean image<=0.0400
Morphological
Drawn Bounding Box

6. Compute Three Local Min
Compute Aspect Ratio
Gait Cycle Detection
Compute Three Local Min
Compute Gait Energy Image
Feature Extraction
Compute Tamura Texture Feature
?train
Recognition No
Yes
NN Classification
Data base
Male/ Female

7. Preprocessing Input Video Frames Morphological Detection
Eigenbackgrounds
If mean image<=0.0400
Morphological
Detection

8. Aspect Ratio =Height/Width (Of Bounding Box).
Gait Cycle Detection
Compute Aspect Ratio:
Aspect Ratio =Height/Width (Of Bounding Box).
Smooth The Aspect:
Smooth the aspect ratio by using moving average filter using a 20-point sliding neighborhood.
Detect The Index:
Detect the index of the first minimum and the index of the third minimum, gait cycle represents the interval that detect between the two indexes.

9. Coarseness/contrast/directionality
Feature Extraction
Tamura Texture Feature
The texture features are coarseness, contrast and directionality from tamura method that extract from gait energy image of one gait cycle.
Gait energy image
Coarseness/contrast/directionality
Feature vector

10. Classification Using Neural Network
Feature vector
Data base
144*3
Cross validation
Train set
120*3
Test set
24*3
Male/Female

11. Classification Result from NN
Classification Accuracy %
Feature type
Data type
Test Data
Train Data 98
100
Tamura texture Features
Male
Female
98%

12. Neural Network Training Performance

13. Neural Network training confusion matrix

14. Conclusions
Gender classification system depended on tamura texture feature is more effective because it depended on gait energy image. This image avoid synchronization difficulties and prevent noises from individual images. This image consider model free gait approach , the model free gait recognition methods or appearance based methods work directly on the gait sequences. They are not considered as a model for the human body to rebuild human walking steps. They have the advantage of low computational cost in compare with model-based approaches and tamura feature considered third ordered texture feature, this type of feature not depended on only pixel but it depended on pixel neighborhoods and this fact give tamura feature strongest to effect by noise for all these reason the system give more robust solution to gender classification.

15. Suggestions for Future Works
The proposed system was working on only side view person walking; in future work suggest dealing with multi view person walking.
suggest dealing with wearing coat and carrying bag CASIS B database and apply the proposed system on CASIS A data base and CASIS C data base.
References

16. THANK YOU …