Newsletter subscription
Please enter your e-mail address to subscribe to our newsletter.                     
 
E-Mail*
Title
First Name
Last Name
Company



The protection of your personal data is very important to us. Please refer to our data protection .

Video fingerprinting and efficient fingerprint matching

There are different approaches to create fingerprints (hashs) of multimedia video data. Different algorithms and techniques shall be evaluated regarding robustness against different manipulations and attacks. It should be tested whether source video fingerprints can be matched to manipulated video fingerprints especially for re-recording with an external camera (e.g. smartphone or webcam) or an extract from source video as partial video copy. The target is to find an efficient fingerprint matching approach to detect re-recording and partial video copies.

Robustness attacks:

  • rotation
  • scaling
  • flipping
  • compression
  • noising
  • cropping
  • color / luminance / chrominance manipulations
  • re-recording with external camera
  • including partial video cuts


Your tasks:

Evaluate fingerprinting technique regarding video data and robustness:

  • Implement the most promising approach as framework
    • best in the language C as library
  • Evaluate your Implementation on different video files , e.g.:
    • execution time
    • true / false positives
    • false / true negatives
    • accuracy

Required skills:

  • Programming: C (preferable, as library), Python, Java, Javascript
  • Video processing: OpenCV, FFmpeg

Related Links:

Supervisor: Sebastian Schmidt

Image fingerprinting and similarity check

There are different approaches to create fingerprints (hashs) of multimedia image data. Different algorithms and techniques shall be evaluated regarding robustness against different manipulations and attacks. It should be tested if source image fingerprints provide a high similarity to manipulated images especially re-recordings with external camera (e.g. smartphone or webcam). Especially interesting are the "perceptual hashing" and "feature matching" approaches applied on a re-recordings.

Robustness attacks:

  • rotation
  • scaling
  • flipping
  • compression
  • noising
  • cropping
  • color / luminance / chrominance manipulations
  • re-recording with external camera


Your tasks:

Evaluate fingerprinting technique regarding image data and robustness:

  • Implement the most promising approach as framework
  • Evaluate your Implementation on different Images (resolution, image content / features, frequencies), e.g.:
  • execution time
  • true / false positives
  • false / true negatives
  • accuracy


Required skills:

  • Programming: C (preferable, as library), Python, Java, Javascript
  • Image processing: OpenCV

Related Links:

    Supervisor: Sebastian Schmidt