Se ha denunciado esta presentación.
Utilizamos tu perfil de LinkedIn y tus datos de actividad para personalizar los anuncios y mostrarte publicidad más relevante. Puedes cambiar tus preferencias de publicidad en cualquier momento.

MediaEval 201 4 Visual Privacy Task: Context - Aware Visual Privacy Protection

264 visualizaciones

Publicado el

In this paper, we describe a privacy filter proposed for VPT 2014 in an attempt to provide a context-aware solution.
The proposed solution comprises three different techniques applied to the face, skin, and body regions separately. The proposed combination of filtering techniques aimed to produce an adaptive solution and provide an example of a context-aware-like privacy filtering capability. The results demonstrated the effectiveness of the proposed techniques in maintaining a high level of Intelligibility
and retaining the appeal of the video i.e. Pleasantness. However the still identifiable gender and race of certain individuals contributed to the perception of lower levels of privacy of the person in the case of some of the video frames.

http://ceur-ws.org/Vol-1263/mediaeval2014_submission_86.pdf

Publicado en: Software
  • Sé el primero en comentar

  • Sé el primero en recomendar esto

MediaEval 201 4 Visual Privacy Task: Context - Aware Visual Privacy Protection

  1. 1. MediaEval 2014 Visual Privacy Task: Context-Aware Visual Privacy Protection Prof. Atta Badii, Ahmed Al-Obaidi School of Systems Engineering University of Reading, UK WWW: http://www.isr.reading.ac.uk email: atta.badii@reading.ac.uk
  2. 2. 2 Introduction • The Proposed Filter - Face filter (H) - Skin filter (M) - Person filter (L) • Evaluation Results - Stream 1 (Crowdsourcing) - Stream 2 (R&D) - Stream 3 (Law enforcement) • Conclusion and future work
  3. 3. 3 Proposed Filter Objectives • Simulate a context-aware system by using a combination of filters applied differently on face, skin, body regions. • Achieve a balance in the well-addressed Privacy- Intelligibility trade-off). •to minimise the potential viewer’s distraction and annoyance caused by the deployed privacy filter.
  4. 4. 4 Proposed Filter 1. Face filter (H) • First an empirical threshold was set to examine the minimum size of the face in which it could be identifiable • Below the said threshold, a simple median blur filter proved sufficient to protect the person’s identity. • Above the threshold, a key point detector was applied on the face region followed by adaptive colour quantisation and circle texturing to produce the final effect.
  5. 5. 5 Proposed Filter 2. Skin filter (M) •colour saturation and luminance values in the R, G, and B colour channels separately modified uniformly within the skin regions. •filtered skin regions were still recognisable. There are only two identifiable skin colours: light-like and dark-like. •skin texture which is responsible for the skin attractiveness was also eliminated.
  6. 6. 6 Proposed Filter 3. Person filter (L) •Edge-based analysis is implemented on the foreground region(s) within the subject bounding box. •morphological operations to enhance the foreground mask to reduce the background noise and minimise the holes in the foreground region. •Canny edge detector is subsequently applied and further refined to eventually draw the subject’s contour. •final effect produced is the result of distance transformation which calculates the distance of each pixel of the resulted binary contour map with the closest zero pixel in the image.
  7. 7. 7 Final Output
  8. 8. 8 Demo 1
  9. 9. 9 Demo 2
  10. 10. 10 Results: Stream 1
  11. 11. 11 Results: Stream 2
  12. 12. 12 Results: Stream 3
  13. 13. 13 Results: All Streams
  14. 14. 14 Conclusion and Future Work • Intelligibility score is well above the region of 70% in the three streams which ensures that the processed video will still serve the main purpose of CCTV security objective. • However, the Privacy scores were slightly below the median which is generally below the value of 50% in this competition. • Regarding the Pleasantness criterion the obtained scores were comparable to the median values for the three streams. • One possible explanation of the overall low Privacy score is the fact that this criterion has been measured based on the ability to identify the gender and the ethnicity of the person which could be hard to conceal without significantly manipulating the person appearance. For instance, the proposed solution would clearly prevent the viewer from being able to identify the featured subject in the normal cases. • Future work/recommendation to include person re-identification test to be included in evaluating the Privacy criterion as an important aspect.
  15. 15. 15 Thank you Atta Badii Intelligent Systems Research Lab (ISR) School of Systems Engineering University of Reading Whiteknights RG6 6AY UK Phone: 00 44 118 378 7842 Fax: 00 44 118 975 1994 atta.badii@reading.ac.uk, www.ISR.reading.ac.uk

×