Loading...

Video Scene Recognition

Diba, Ali | 2013

1645 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 44757 (19)
  4. University: Sharif University of Technology
  5. Department: Computer Engineering
  6. Advisor(s): Ghanbari, Mohammad
  7. Abstract:
  8. Scene classification and understanding is one of the most important fields in computer vision. Its applications are such as exploring robot navigation enviroment, content-based image retrieval (CBIR), organization in image databases, highly semantic describing images and videos and content extraction of videos.Many methods and algorithm are proposed till today to deal with diversity of this field by emphesizing on feature based methods or machine learning based methods. In this research we have focoused on proposing a new algorithm which is using principals of NBNN image classification method but major changes in how to exract distance metric from Nearest neighbour and how to use local features and descriptors like SIFT, HOG, Gist ... for computing these distance. We use our new method to classify video events by segmenting video frames to important objects and main parts for extraction more accurate visual dictionary of video sources. Hiring state of the art algorithm like max-margin multiple instance learning helped us to achieve more reliable visual dictionary for each category of scenes or events.This method gain state of the art results rescpet to the other related methods including parametrics and non-parametrics on the famous indoor and outdoor scene databse like MIT 67-indoor scene, 15 category scene, UIUC 8 sports events and video event MED11 dataset
  9. Keywords:
  10. Classification ; Scene Recognition ; Nearest Neighbor ; Scene Classification ; Video Images ; Image Segmentation ; Naive Bayes Nearest Neighbor (NBNN)

 Digital Object List

 Bookmark

...see more