1 / 11

Facial Feature Detection

Facial Feature Detection. Levente Sajó University of Debrecen. Human Computer Interaction. In multi-modal human-computer interaction takes an important part face detection/recognition extracting facial features emotion detection age recognition. Face Detection.

Download Presentation

Facial Feature Detection

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Facial Feature Detection LeventeSajó University of Debrecen

  2. Human Computer Interaction • In multi-modal human-computer interaction takes an important part • face detection/recognition • extracting facial features • emotion detection • age recognition

  3. Face Detection • For detecting faces, many different techniques appeared over the years • Template based • Appearance based (neural networks, SVM) • Probably the most successful is the one based on cascaded Haar-classifiers (Boosted Cascade Detector - BCD) • On the localized face further steps can be performed for recognizing gender, age or facial gestures

  4. Emotion Detection • 6 different facial emotions: neutral, happy, sad, surprised, angry, fear, disgust • Classification methods used in face detection can be used for emotion detection, too: • Gabor-transformed image is classified using SVM or BCD • A feature vector formed by manually defined facial landmarks is passed to SVM classifier

  5. Emotion Detection • Emotion detection is sensitive for changes of illumination and different rotation of the face • Using 2 cameras, 3D feature points can be used for constructing the feature vectors, with these more accurate classifiers can be created

  6. Localizing Facial Features • Local feature detectors (SVM, BCD) can be used to detect facial features • Since facial features contains less information then the whole face, individual feature detectors seemed to be unreliable

  7. Localizing Facial Features • Shape models can be used to • reduce the number of false detections by only selecting plausible configurations of feature matches • correcting the false detection of the local feature detectors • Statistical Shape Model • For each landmarks their mean position and variance are determined • Distance Shape Template

  8. Distance Template • The template is described by template rules • A rule defines the estimated distance between template points • If a template point does not satisfy the conditions of a rule, a penalty value is calculated • The sum of the penalties gives the overall penalty of the template

  9. Distance Template • By replacing the feature points, the overall penalty of the template can be minimized

  10. Conclusion • Emotion detection is a complex task • Single techniques proved to have several weaknesses • Combination of techniques can result a robust emotion detection

  11. Thanks for attention!

More Related