1 / 12

Presented by Hugo Groening

ECE 172A SIMPLE OBJECT DETECTOR WITH INDICATOR WHEN A NEW OBJECT HAS BEEN ADDED TO OR MISSING IN A ROOM. Presented by Hugo Groening. INTRODUCTION.

elias
Download Presentation

Presented by Hugo Groening

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. ECE 172ASIMPLE OBJECT DETECTOR WITH INDICATOR WHEN A NEW OBJECT HAS BEEN ADDED TO OR MISSING IN A ROOM Presented by Hugo Groening

  2. INTRODUCTION In the security industry, there has always been the demand for simple, accurate and economical systems that allow the prevention or ability to track through footage when a security breach has occurred. The project to be presented targets locations such as museums, stores, homes, Banks, etc.

  3. AGENDA • Objective • Related Research • Method • Results • Future work or improvements • Conclusion

  4. Objective • To create an object recognition and detector in a room using AVI files. • To train detector to indicate when an object is missing or possibly about to be taken.

  5. Related Work and Research Industry - Geovision: http://www.geovision.com.tw/demo/vid/demo4.htm References: - A. Torralba, K. P. Murphy and W. T. Freeman. (2004). "Sharing features: efficient boosting procedures for multiclass object detection". Proceedings of the 2004 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR). pp 762- 769 - B. C. Russell, A. Torralba, K. P. Murphy, and W. T. Freeman. Labelme: a database and web-based tool for image annotation. Technical Report AIM-2005-025, MIT AI Lab Memo, September, 2005. Can be found at: http://people.csail.mit.edu/brussell/research/AIM-2005-025-new.pdf Location of Code and Database: http://people.csail.mit.edu/torralba/shortCourseRLOC/boosting/boosting.html

  6. Goal Edited from http://www.geovision.com.tw/demo/vid/demo4.htm, Geovision Inc.

  7. Initialize Video output Choose Video to be analyzed Look at 1st frame to find size Detect if object count changes from previous frame loops into framesand makes copy of current frame Nothing change, keeps counting Count Objects in frame If different Change to Grey and invert to BW Erodes, dilates WARNING Find centroid of objects and tracks them Method

  8. Results • Count Objects • Find Objects • Warns when object count changes • Storages Video Output for Future analysis

  9. Limitations • Hard drive Space for memory allocation of Database • Processor • AVI files taken from other video cameras • Objects and background color • Time frame

  10. Output Footage

  11. Future Work or Improvements • Minimize the memory size of Database • Funding for continuation of Labelme Database • Access to individual Objects within Database • Research other Algorithm options for Object detection and Classification • Find more adequate video Recorder for surveillance purposes • Live Object tracking detector • To become more knowledgeable with Matlab and other Programming languages.

  12. Conclusion The background and knowledge obtained through the practice of the information previously researched, had developed a strong awareness of how digital image processing can be directly applied to needs in society. In even the most simple cases, such as in the security or surveillance industry, high end technology will be needed for accuracy. Continuous development and improvements will also be necessary.

More Related