sign language recognition using HMM

2,226
-1

Published on

1 Comment
1 Like
Statistics
Notes
  • This is amazing!!!
    I am looking to do exactly the same in my graduation project!
    Can you help me and provide me with the source code, please?
       Reply 
    Are you sure you want to  Yes  No
    Your message goes here
No Downloads
Views
Total Views
2,226
On Slideshare
0
From Embeds
0
Number of Embeds
0
Actions
Shares
0
Downloads
129
Comments
1
Likes
1
Embeds 0
No embeds

No notes for slide

sign language recognition using HMM

  1. 1. Sign Language RecognitionUsingHidden Markov Model<br />Presented by:<br />VipulAgarwal - 070905060<br />
  2. 2. Outline<br /><ul><li>INTRODUCTION
  3. 3. SIGN LANGUAGE
  4. 4. PRE-PROCESSING
  5. 5. SKIN AND HAND DETECTION
  6. 6. OPTICAL FLOW ANALYSIS
  7. 7. FEATURE EXTRACTION FOR TRAINING DATA
  8. 8. HIDDEN MARKOV MODEL & ITS USE
  9. 9. PROGRESS REPORT
  10. 10. DEMONSTRATION</li></li></ul><li>Introduction<br /><ul><li>Interaction with computers may often not be a comfortable experience.
  11. 11. Computers should be able to communicate with people with body language.
  12. 12. Hand gesture recognition becomes important …</li></ul>Interactive human-machine interface and virtual environment <br />
  13. 13. Introduction<br />Two common technologies for hand gesture recognition<br />GLOVE-BASED METHOD<br />Using special glove-based device to extract hand posture<br />VISION-BASED METHOD<br />3D hand/arm modeling<br />Appearance modeling<br />
  14. 14. Introduction<br />3D hand/arm modeling<br />Highly computational complexity <br />Using many approximation process<br />Appearance modeling<br />Low computational complexity<br />Real-time processing<br />
  15. 15. Sign Language<br />Rely on the hearing society<br />Two main elements:<br />Low and simple level signed alphabet, mimics the letters of the spoken language.<br />Higher level signed language, using actions to mimic the meaning or description of the sign.<br />The project aim is to make the computer recognize low and simple level American Sign Language.<br />
  16. 16. Sign Language<br />American Sign Language<br />26 signs to denote the alphabets.<br />10 signs to denote numbers<br />
  17. 17. Pre - Processing<br />The video sequence used has a lot of noise due to:<br />Low quality of the webcam <br />Improper lighting conditions<br />Background<br />
  18. 18. Pre - Processing<br />Pre-processing involves reducing the noise and illumination problems.<br />The morphological operations used for reducing the noise involves:<br />Dilation<br />Statistical Elimination<br />
  19. 19. Pre - Processing<br />DILATION><br />A disc shaped region is traversed over every blob and the ones which do not fit the disc are removed completely.<br />
  20. 20. Pre - Processing<br />STATISTICAL ELIMINATION><br />For every region the area is computed. Since hand is the one with the largest area, all blobs having less than a specified area are removed.<br />
  21. 21. Hand Detection<br />First all the noise is removed in the pre-processing stage.<br />Now we assume that the hand is the largest skin blob in our video sequence.<br />We calculate the area of every blob and take the one with the largest area.<br />We also calculate the bounding box of the region containing the hand for further analysis<br />
  22. 22. Hand Detection<br />
  23. 23. Optical Flow Analysis<br />DEFINITION:<br />Optical flow is the pattern of apparent motion of objects, surfaces, and edges in a visual scene caused by the relative motion between an observer (an eye or a camera) and the scene.<br />
  24. 24. Optical Flow Analysis<br />Why Optical Flow Analysis?<br />Till now the system is just able to detect the hand and follow the bounding box as the hand moves.<br />The problem now is that we need to define a way to take a snapshot of the hand when the hand is not moving.<br />
  25. 25. Optical Flow Analysis<br />Using this technique we find the motion in the hand. When the hand has stabilized, we assume that the gesture is ready. We then take a snapshot of the hand and perform the recognition on that image.<br />
  26. 26. Feature Extraction<br />For training the network with test images we perform the following feature extraction technique:-<br />Thresholding of the test hand<br />Converting to a binary image<br />Finding the centroid of the hand and orientation of the minor axis.<br />Making feature vectors using a predefined number of features.<br />
  27. 27. Feature Extraction<br />Extracting the intersection of the feature vectors with the boundary points.<br />Finding the scalar length of the vectors from the centroid.<br />Normalising the lengths in a scale of 1 to 100 to make it scaling invariant.<br />
  28. 28. Feature Extraction<br />
  29. 29. Hidden Markov Model (HMM)<br />HMMs allow you to estimate probabilities of unobserved events<br />Given plain text, which underlying parameters generated the surface<br />
  30. 30. HMMs and their Usage<br />HMMs are very common in Computational Linguistics:<br />GESTURE RECOGNITION (observed: image, hidden: alphabets)<br />
  31. 31. Progress Report<br />WORK COMPLETED:<br />Data Collection<br />Pre-processing <br />Skin And Hand Detection<br />Optical Flow Analysis<br />Feature Extraction For Training Data<br />WORK REMAINING:<br />Training The Hidden Markov Model<br />
  32. 32. Any Questions …?<br />

×