Loading presentation...

Present Remotely

Send the link below via email or IM

Copy

Present to your audience

Start remote presentation

  • Invited audience members will follow you as you navigate and present
  • People invited to a presentation do not need a Prezi account
  • This link expires 10 minutes after you close the presentation
  • A maximum of 30 users can follow your presentation
  • Learn more about this feature in our knowledge base article

Do you really want to delete this prezi?

Neither you, nor the coeditors you shared it with will be able to recover it again.

DeleteCancel

Make your likes visible on Facebook?

Connect your Facebook account to Prezi and let your likes appear on your timeline.
You can change this under Settings & Account at any time.

No, thanks

Lip Reading

LipReading is a final project of Ben Gurion University Software Engineering. The system identifies the user’s lip movements by video, and transforms the video to audio output of the spoken word.
by

Dor Leitman

on 25 June 2013

Comments (0)

Please log in to add your comment.

Report abuse

Transcript of Lip Reading

Lip Reading
What is
Lip Reading System?
Technologies
System Architecture
Stand alone mode
Client-Server mode:
Results
How Does It Work?
Given a video segment, an image-processing algorithm identifies the lips and extracts the coordinates of a number of points on the lips from each frame.
Auto Lip Recognition Algorithm:
For each frame:
Identify lip area using Haar function (OpenCV)
Finding middle upper and lower lip coordiantes
Finding the edges of the lips
Cons: Requires certain light conditions
User put stickers on his lips and configure the colors
Stickers colors are saved in application
For each frame:
Finding the pixels in stickers colors (OpenCV)
Cons: Requires stickers on user lip
Image Processing
Colored Stickers Recognition Algorithm:
Stretch Normalizer
Rotation Normalizer
Center Normalizer
Time Normalizer
Resolution Normalizer
Skipped Frames Normalizer
Multi Layer Perceptron Classifier
Support Vector Machine Classifier
Dynamic Time Wrapping Classifier
Classification
Normalization Of The Data:
Lip coordinates data is going through a series of normalization actions to improve the classification results.
Using Weka Machine Learning algorithms package
Classification:
Classification model containing data of many records of words/phrases
Classifier returns a corresponding string of the new data instance according to the model.
Result
String returned from Classifier to voice
using Google Voice Library
Lip Reading Application
Communication for everyone...
The End
Supervisors:
Dr. Kobi Gal
Dr. Gavi Kohlberg

Sagi Bernstein
Dor Leitman
Dagan Sandler
Full transcript