Using semantics to analyze body camera video could help law enforcement agencies better manage the large amounts of video data. The semantic analysis can extract text, metadata and identify key ideas and words to allow agencies to more easily search and query large volumes of video. However, fully realizing this solution will require integrating multiple systems for video storage, semantic analysis, and query capabilities to address challenges such as privacy regulations and budget constraints across different jurisdictions.

1. Using Semantics to Enhance
Bodyworn Camera Video for Law Enforcement

2. Mike Doane
Lecturer
Information School
University of
Washington
Gravity Zero
Director,
Information Services
ABOUT ME

3. Effective Solutions for
Managing Police Body Camera
Video

4. Takeaway Ideas
o Law enforcement needs a better way to
manage body camera video
o Using semantics to help analyze video can
reduce effort and improve usage
o Solution requires integrating multiple types
of systems and capabilities

5. 10
%
What is a Body Worn Video
Camera?

6. 10
%
How Effective are Body
Video Cameras?

7. What’s the Problem?

8. What’s the Problem?
As more police agencies equip
officers with body cameras in
response to public pressure,
authorities are discovering they
create problems of their own:
how to analyze, process and
store the mountains of video
each camera generates.

9. What’s the Problem?
Right now, there are too many
body camera videos being
produced for law enforcement
agencies to effectively analyze

10. What’s the Problem?
This problem is being
compounded daily by more
cameras producing more videos
that are simply being stored to a
cloud system

11. What’s the Problem?
The cameras create an
enormous amount of data.
Who gets access to it? How
does it get stored? What
appeared to be a no-brainer in
terms of bringing accountability
to the force has raised a lot of
ancillary questions.

12. How Many Police Agencies
are There in the US?
765,000
officers
18,000
agencies
+5%

13. Who’s Got the Cameras?
4,500
agencies= have
13,500
agencies= want
Only 25%
currently
have cameras

14. How Much Storage is
Needed?
110
officers
8,000 – 10,000
videos /month
US$2.6M
year
Evidentia
ry vs.
Non-
Evidentia
ry
Example
using one
police
departme
nt
Estimat
ed cost
of
storag
e for
the
first
year

15. Our Areas of Research
Video
management
Semantic
Extraction
Standardized
vocabularies
How can
agencies
determine
if there
is
relativel
y
important
evidence
on a
How
should
video
be
create
d and
stored
?
Can we
create
a
unified
ontolo
gy for
all
levels
of law
enforc

16. What’s the Solution?

17. What’s the Solution?
Upload
and
manag
e
videos
Analyz
e text
for
seman
tic
meanin
g
Automati
cally
extract
text and
metadat
a from
videos
Query to
discover
key ideas
and
specific
words
within
videos
?

18. What Does Each Provide?
Provid
es the
online
video
storag
e and
proces
sing,
plus
advanc
ed
video
manag
ement

19. What Does Each Provide?
Provid
es the
video
to
text
extrac
tion
servic
e plus
keywo
rd
taggin
g
capabi
lity

20. What Does Each Provide?
Provides
the
semantic
analysis
capabilit
y plus
the
ontology

21. What Does Each Provide?
Provid
es the
servic
es and
integr
ation
capabi
lity

22. Challenges Ahead
Scalability
REGULATION Privacy
50 states
have
different
regulatio
ns
regarding
video
usage
Most
police
agencie
s are
small
to
midsize
and
have
limited
budget
s
Just
becaus
e video
can be
used,
who
gets to

23. Our Next Steps
Proof of
Concept
Scaled Testing
Overcome
Doubt
Need to
validate
how
large the
solution
can be
Working
with a
few
agencie
s now
for PoC
Police
are
natura
lly
wary
of
being
record

24. Mike Doane
mtd@uw.edu
mike.doane@gravityze
ro.com
www.linkedin.com/in/mikedoane
CONTACT ME

25. appendix
Slide
3
Source
https://www.bostonglobe.com/metro/2015/08/05/boston-city-council-examine-bodycam-proposal-
hearing/1Lk3a6GTQ1qkmkrPK1LrgO/story.html
4 http://gizmodo.com/seattle-police-put-redacted-body-cam-footage-on-youtube-1689139204
5 http://www.visualnews.com/2015/08/11/we-look-at-how-police-body-cameras-work-and-how-effective-they-are/
6 http://www.visualnews.com/2015/08/11/we-look-at-how-police-body-cameras-work-and-how-effective-they-are/
7 http://blackathlete.net/2014/08/state-america-suburbs-st-louis/
8 Zusha Elinson and Dan Frosch, Wall Street Journal, April 9, 2015
9 Laurie Capitelli, Berkeley (CA) City Council member. Wall Street Journal, April 9, 2015
11 http://www.policeone.com/police-products/body-cameras/articles/8243271-For-police-body-cameras-big-costs-loom-
in-storage/
12 https://en.wikipedia.org/wiki/Law_enforcement_in_the_United_States#Number_of_police

26. appendix
Slide
13
Source
http://www.nij.gov/topics/law-enforcement/technology/pages/body-worn-cameras.aspx
14 http://news.yahoo.com/police-body-cameras-big-costs-loom-storing-footage-174207101.html
17 https://www.oracle.com/content-storage-management/feature.html
18 http://www.ramp.com/feature/metadata-transcription-tagging/
19 https://www.poolparty.biz/