MedDMO multimedia tools for countering disinformation

The MedDMO Multimedia Analysis Toolbox for countering disinformation

1. This project has received funding from the European Union
DIGITAL-2021-TRUST-01. Contract number: 101083756
meddmo.eu
The MedDMO Multimedia Analysis Toolbox
for countering disinformation
Manos Schinas, Panagiotis Galopoulos, Efthimios Hamilos, Nikos Sarris,
Ioannis Kompatsiaris and Symeon Papadopoulos
Nikos Sarris
Media analysis, verification and retrieval group (MeVer)
Center for Research and Technology Hellas (CERTH)

2. Objects and actions
are automatically
identified and
added as tags
Automatic image
analysis creates
short descriptions
Assets are
classified to
specific categories,
as AI-generated,
disturbing, NSFW
or memes
Multimedia Archive
A platform where you can easily analyse, find and annotate media
Any tag can be
manually added by
users

3. Disturbing and
NSFW images are
redrawn to conceal
disturbing/insulting
details
Any of the three
types of images can
be filtered in or out
Disturbing, NSFW
(not safe for work)
or memes are
automatically
classified
Automatic classification
Specific types of content are automatically identified and blurred
Applies also to
manually tagged
content

4. Visual heatmaps
help users quickly
understand the
results
15 image analysis
algorithms help
uncover possible
forgeries
Heatmap overlay
on image can help
users spot
manipulated areas
Image Verification Assistant
Quickly analyse images to identify possible forgeries
Manual annotation
space helps users
discuss findings

5. Temporal video
segmentation
illustrating the
manipulation
probability of
every segment
Top level analysis
for the entire
video
Image/video
player window for
detailed viewing
Deepfake detection
Assess the possibility of deepfake manipulation in image or video
Detailed analysis
per video segment

6. Probability
shown along with
suggested
outcome tagged
Several models
used for detection
of various
generation
technologies
Image under
question
Synthetic image detection
Assess the possibility of image having been synthetically generated
Detailed
information for
every model
provided

7. Map illustrating
the most likely
location for the
image
Window for
detailed image
viewing
Location estimation
Estimate the most likely location for an image based on visual cues
Example images
from the most likely
location illustrating
visual similarity
with the query
image

8. Results include
images of semantic
similarity to the
query image
Image for which
similar ones are
sought
Search by visual similarity
Find images or videos that are visually and semantically similar
Selecting to search
for similar images

9. Results include
images that are
almost (visually)
identical to the
query image
Search by visual similarity
Find images or videos that are visually nearly identical
Selecting to search
for near duplicate
images

10. CONTACT
us
This project has received funding from the European Union
DIGITAL-2021-TRUST-01. Contract number: 101083756
meddmo.eu
Fact-Check by MedDMO
@MEDDMOhub
[email protected]