In a near future, a greater number of individuals in long-term care will live alone. New solutions are needed in order to provide them support and increase their autonomy at home. Intelligent monitoring systems based on computer vision may provide a solution. However, privacy related issues must be solved beforehand. In this paper, we propose a level-based visualisation scheme to give users control about their privacy in those cases in which another person is watching the video. These visualisation levels are dynamically selected according to the context by displaying modified images in which sensitive areas are protected.
Paper at http://link.springer.com/chapter/10.1007/978-3-319-13102-3_55
3. 3
Why (visual) privacy?
AAL services provide help to
people in need of long-term
care.
Development of novel video-based
AAL services for private
spaces.
Some AAL services may require
human access to video
stream.
...
Camera 1 Camera 2 Camera N
...
...
Long-term
analysis
Setup and Profiles
DB (Activities,
Inhabitants,
Objects, ...)
Log
Event
Alarm Actuators
Caregiver
Motion
Detection
Motion
Detection
Motion
Detection
Human
Behaviour
Analysis
Human
Behaviour
Analysis
Human
Behaviour
Analysis
Multi-view Human Behaviour Analysis
Environmental
Sensor Information
Reasoning
System
Privacy
Architecture of our Intelligent Monitoring System
4. 4
but... What is privacy?
Privacy: Sphere of the private life that an individual has right
to protect from intruders.
However...
The notion of privacy is highly subjective. It depends on the
individual.
Several factors are involved:
● The private “thing”
● Observer / intruder
5. 5
The private “thing”
“A picture is worth a thousand words”
An image conveys the following information about individuals:
● Identity (Who?)
● Appearance (How?)
● Location (Where?)
● Activity / Behaviour (What?)
● Time (When?)
6. Image Redaction: Modify an image or a sequence of
images so as to protect objects (visual clues) appearing on
them.
6
How to ensure (visual) privacy?
But...
● Image must retain its utility
● A trade-off between privacy protection and image utility is
needed
● Privacy must be adaptable to the individual
7. 7
Privacy by Context
● We propose a privacy
protection scheme that is
aware of the context
● A set of redaction methods
is used
● A context describes “any”
situation.
● Users provide their privacy
preferences by linking
instances of the context with
protection methods
Blur
Pixel
Emboss
Naïve
Skeleton Avatar 3D
Silhouette
Invisibility Selected Method
Skeleton
Set of redaction
methods
User-given
Matching
Context
Composed of several
variables
8. 8
Used Context
● The observer / viewer / watcher
● Identity of the person (to retrieve the privacy profile)
● Closeness between person and observer (e.g. relative,
doctor, acquaintance, … )
● Appearance (dressed?)
● Location (e.g. kitchen)
● Event (e.g. cooking, watching TV, fall, ...)
10. 10
Conclusions
● Visual Privacy by Context covers several research fields
Computer vision (HBA, Object detection, Person Re-id, …), Image processing and
in-painting, AI, Sociology.
● As privacy is highly subjective, any solution should be
adaptable to each individual
● Need of a privacy measure so as to objectively evaluate
protection methods