Presentation is loading. Please wait.

Presentation is loading. Please wait.

CMPT-884 Jan 18, 2010 Video Copy Detection using Hadoop Presented by: Cameron Harvey Naghmeh Khodabakhshi CMPT 820 December 2, 2010.

Similar presentations


Presentation on theme: "CMPT-884 Jan 18, 2010 Video Copy Detection using Hadoop Presented by: Cameron Harvey Naghmeh Khodabakhshi CMPT 820 December 2, 2010."— Presentation transcript:

1 CMPT-884 Jan 18, 2010 Video Copy Detection using Hadoop Presented by: Cameron Harvey Naghmeh Khodabakhshi CMPT 820 December 2, 2010

2 Introduction  Video Copy Detection is an important tool for detecting copyright infringements  A copy can be obtained through a set of transformations from the original such as:  Video cropping or scaling  Gamma shift  Blurring  Addition of logo  Changes in quality (noise, framerate, …)

3 Copy Detection Process  There are 2 stages in the Copy Detection Process  Signature Extraction The video content is used to create a unique signature of the reference and the query video  Signature matching A distance metric compares the signatures of the videos If they are close enough, the query is considered to be a copy

4 SURF - Speeded up Robust Features  A method to find and extract a set of interest points from an image  It is based on SIFT – Scale Invariant Feature Transformation

5 Our Solution  We implemented the Copy Detection algorithm of Roth et. al. [1] but parallelized the process to speed up processing  For each frame in the video we divide the image into 4x4 giving 16 regions [1] G. Roth, R. Lagani`ere, P. Lambert, I. Lakhmiri, and T. Janati. A simple but effective approach to video copy detection. In CRV ’10: Proceedings of the 2010 Canadian Conference on Computer and Robot Vision, pages 63–70, Washington, DC, USA, 2010. IEEE Computer Society.

6 Our Solution (2)  A signature for each frame is created by counting the number of SURF features discovered in each region. The signature is a 16-dimensional vector  We compare frames using a distance metric  If the distance is below a threshold, then the frames are considered to be a match

7 Our Solution (3)  A video is considered a copy if there are a significant number of consecutive matching frames

8 Hadoop: Map-Reduce  Based on key-value data structures  MAP  The input to the MAP function is a list of key-value pairs.  A user defined function is applied to every element of the list  REDUCE  The output of the Map is sorted based on the keys and passed to the Reduce function  The Reduce function joins values with the same keys

9 Parallelizing the Process

10 Parallelizing the Process (2)

11 Results: Gamma Correction

12 Results: Scaling

13 Results: Blurring

14 Results: Hadoop  Signature extraction using a single node  1 hour 2 minutes and 11 seconds  Signature extraction using 6 nodes  12 minutes and 3 seconds

15 Questions


Download ppt "CMPT-884 Jan 18, 2010 Video Copy Detection using Hadoop Presented by: Cameron Harvey Naghmeh Khodabakhshi CMPT 820 December 2, 2010."

Similar presentations


Ads by Google