Ref. Ares(2018)4699176 - 13/09/2018
PROJECT SIIP
Deliverable 9.4:
Field Testing and End-User Training Final Report
Date of publication: 30 April 2018
Reference: LA/51275-54/5.4/JBA/XB/tsa
Due date of deliverable: 30 April 2018
Organization name of lead contractor for this deliverable: INTERPOL
Dissemination level: Restricted to a group specified by the consortium (including the Commission Services)
Start date of project: 1 May 2014
Duration: four years
link to page 3 link to page 5 link to page 5 link to page 5 link to page 6 link to page 8 link to page 8 link to page 8 link to page 8 link to page 9 link to page 9 link to page 10 link to page 10 link to page 10 link to page 11 link to page 11 link to page 12 link to page 16 link to page 17
Deliverable D9.4: Field Testing and End-User Training Final Report
TABLE OF CONTENT
1
INTRODUCTION ....................................................................................... 3
2
OVERVIEW .............................................................................................. 5
3
PARTICIPANTS ......................................................................................... 5
3.1
Consortium Meeting ............................................................................... 5
3.2
Field Test and End-User Training ............................................................. 6
4
DISCUSSIONS .......................................................................................... 8
4.1
Key note speech ...................................................................................... 8
4.2
SIIP overall presentation ......................................................................... 8
4.3
Demonstration ........................................................................................ 8
4.4
Enrichment presentations ....................................................................... 9
4.4.1
German Federal Police .............................................................................................. 9
4.4.2
International Biometric Group ................................................................................. 10
4.4.3
New Scotland Yard .................................................................................................. 10
4.4.4
Portuguese Criminal Police ...................................................................................... 10
4.4.5
Netherlands Forensics Institute ............................................................................... 11
5
END-USER TRAINING ............................................................................. 11
6
DISSEMINATION .................................................................................... 12
7
CONCLUSION ......................................................................................... 16
ANNEXES .......................................................................................................... 17
Page 2/23
Deliverable D9.4: Field Testing and End-User Training Final Report
1 INTRODUCTION
The present document consists of deliverable D9.4 “Field Testing and End-User Training final report” of
the Speaker Identification Integrated Project (SIIP). SIIP is a European Union funded research project,
under the 7th Framework Program (FP7), aimed at developing law enforcement capabilities in the field of
voice recognition and speaker identification, including the ability for law enforcement agencies (LEA) in
one country to search databases of speakers in other countries via INTERPOL. Law enforcement agencies
in INTERPOL member countries and INTERPOL itself represent the potential end-users of the SIIP
outcome.
Deliverable D9.4 “Field Testing and End-User Training final report” is prepared within the scope of Work
Package 9 “Validation, Field Testing and End Users training” efforts as detailed in
Box 11. The report
starts by presenting the objectives and structure of the event hosted by INTERPOL at its General
Secretariat premises on 21-22 November 2017. Further on, it describes the event audience, presenting
the participants’ background. The report contains a detailed outline of the event content, including the
live demonstrations performed by two LEAs in their capacity as project partners, the enrichment
presentations, the training session and adjacent discussions. Additionally, the deliverable expands on the
dissemination efforts undertaken prior to, during and following the event, including the associated
material. The report ends with some conclusions and final thoughts on the overall impression of the end-
user community regarding the developed tool. For readers’ convenience, the annexes comprise the
event agenda, template invitation letter circulated to the stakeholders, group photo as well as the
updated version of the SIIP factsheet.
To contextualise the endeavours presented in the present deliverable,
Box 1 below reproduces from the
SIIP Description of Work (DoW), the Work Package 9 objectives, the tasks of relevance to the report, and
the expected deliverables.
Box 1. Extract from SIIP Description of Work: Work Package 9 “Validation, Field Testing and End Users training” (with minor
edits)
Objectives
A. Validation
- to validate and quantify the real-world performance of the integrated SIIP solution, with a focus on quantifying
improvements over current-generation voice identification technologies;
- to identify the factors that most directly influence performance in text-independent voice identification systems, such as
input device, operational environment, subject composition;
- to provide the EC with information required to substantiate and justify the use of SIIP technologies in a range of
environments central to the security domain.
B. Field Testing and End-Users training
- to run the field tests to test the integrated solution in real field environment and scenarios;
- to train SIIP end users so they will be able to operate SIIP system directly during the field tests;
- to train INETRPOL, PJ, MPS, BKA and RACIS investigation officers about SIIP/SISC portal & MMI operation, features and
capabilities (so they will be able to operate the system directly during the field tests).
Description of work
T9.4 Field Test Plan will run, try and validate the SIIP eco-system (SIIP<-->SISC) in 3 theatres in parallel (PJ, RACIS and
INTERPOL) in order to demo the SIIP eco-system capabilities. This will ensure that the developed solutions are feasible in a
variety of LEA investigation environments.
The primary objective of the field test tasks will be to allow the consortium to
1
Box 1 reproduces from the SIIP Description of Work, the WP9 objectives, the description of tasks of relevance to the present
report, as well as the list of expected deliverables.
Page 3/23
Deliverable D9.4: Field Testing and End-User Training Final Report
install, commission and conduct the various tests required to prove the effectiveness of the SIIP solution.
Therefore the Field Test Plan will have to take into account the following points to ensure its success:
• Provision of sufficient space to install the SIIP equipment, provision of power for the SIIP equipment, provision of Local
Area Network (LAN) connectivity, hosting of any IT equipment (servers etc.) on campus that may be required, WAN
connectivity required to access components of the SIIP solution;
• Provision of LEA staff to facilitate field tests (by prior arrangement);
• Co-ordination with any person and entity that are involved in the field tests;
• Scenarios content and scenario acting procedures and personnel;
• Preparation of scenarios simulated data;
• The way of describing and the scenario story to the audience;
• Event dissemination plan, event agenda, event guests/audience list , invitations , event documentary movie;
• Field test day logistics (catering, seats, accommodation etc.);
• Clear definition of the test strategy and test plans.
T9.5 Field Tests Execution
For the purpose of executing field tests for the SIIP solution the following sub-tasks will be carried out:
• INTERPOL will provide lab space where SISC will be installed; PJ will provide lab space where the SIIP (PJ) will be installed;
RACIS will provide lab space where the SIIP (RACIS) will be installed.
• The consortium will act and simulate two field test scenarios:
1. Scenario #1: Jewelleries robbery investigation
2. Scenario #2: Terrorist attack investigation
• Perform two Field Test Demo day-events, one at INTERPOL and second at PJ. These two events will be a show of SIIP
consortium where the system SID capabilities and international info sharing mechanism will be demonstrated (based in
simulated data) to external guests (LEAs and policy makers from all Europe). Therefore, this event can be also considered as
dissemination event.
• During the field test event day, the consortium will simulate each of these scenarios by acting them, step by step, and
show to the guests the system capabilities and performance.
• The whole SIIP eco-system (SIIP-SISC) will be activated and demonstrated including:
o SIIP/SISC portal (which will be operated directly by PJ, RACIS and INTERPOL officers who will be trained for
operating the system portal and MMI (T9.6);
o SIIP and SISC functionalities;
o SIIP secured, privacy preserving model, for international info sharing of voiceprints and metadata.
T9.6 Field Tests results analysis and development of recommendations and findings on extensibility of technology to
security domain
Results from SIIP assessment and validation will support development of recommendations and findings on the
extensibility of the SIIP technology to security domain. These recommendations will explore several dimensions of
extensibility, including scale (number of enrolments, range of environments in which cross-platform identification is
possible); optimizing trade-offs between automated and analyst-driven identification to support highly reliable
identification; and determining the minimum voice sample that can be used to support reliable identification.
Description of deliverables
D9.1) Validation and Performance Testing Methodology
D9.2) Validation and Performance Testing In-Progress POC Report
D9.3) Validation and Performance Testing Final Report and Recommendations
D9.4) Field Testing and End-User Training Final Report
Page 4/23
Deliverable D9.4: Field Testing and End-User Training Final Report
2 OVERVIEW
From the 21st to the 22nd of November 2017, INTERPOL hosted the second and final SIIP Field Test & End-
User Training at its General Secretariat premises in Lyon, France. The event’s main objective was to
validate the SIIP system’s operational capabilities using a research database with real audio-recordings
and to organize a set of workshops dedicated to training end-users in handling the SIIP system.
Furthermore, the end-user feedback would contribute to enhance the technical tools and individual
capabilities, improve the user interface portals and refine the different evaluation criteria.
The two-day event was organised as follows:
First, the Consortium Meeting was held on 21 November. It was attended by a smaller audience
comprising project partners involved in the execution of the Field Test. This rehearsal meeting
had the objective to ensure the smooth organisation and execution of the final Field Test the
following day by testing the equipment, ensuring the availability of necessary connection and
infrastructure, fine-tuning any last-minute needs and demands and finalising coordination
between the actors;
Followed by the Field Test and End-User Training organised on 22 November and attended by a
large audience consistent of consortium partners, numerous external participants, INTERPOL
staff from different departments and European Commission representatives.
A social event was organised at the end of the rehearsal meeting, allowing the partners to get together
and welcome the experts that committed to attend the final Field Test the following day. The social
programme included a sightseeing boat tour combined with dinner, enabling the participants to engage
in discussions and exchange opinions about the SIIP project in an informal setting.
3 PARTICIPANTS
3.1 Consortium Meeting
The first day rehearsal meeting was attended by the following consortium partners:
Airbus Defence and Space
Carabinieri, Italy
Data Fusion International
Idiap Research Institute
INOV - INESC Inovação Instituto de Novas Tecnologias
INTERPOL
NUANCE
Laboratory of Citizenship Sciences (Laboratorio di Scienze della Cittadinanza – LSC)
Metropolitan Police (MET) UK
Novetta
OK2GO
Page 5/23
Deliverable D9.4: Field Testing and End-User Training Final Report
University of Groningen
SAIL Labs Technology
SINGULARLOGIC
Synthema
Verint
Federal Criminal Police Office of Germany (Bundeskriminalamt - BKA)
Portuguese Criminal Police (Policia Judiciaria - PJ)
Warwick University
3.2 Field Test and End-User Training
The Final Field Test & End-User Training was attended by 110 participants representing more than 50
countries coming from a diverse range of backgrounds, with different but mutually complementary areas
of expertise.
The attendees included representatives from the following categories:
SIIP consortium members;
External experts invited by INTERPOL, including:
o Officers from over 40 law enforcement authorities (the BKA, Swiss Police, the General
Department of Criminal Evidence of Kuwait, NCB2 India, NCB Guinea-Bissau etc.);
o Several INTERPOL directorates (i.e. Fugitive Investigative Support, Maritime Security,
Counter-Terrorism, Information Security);
o Representatives of academia and the private sector (i.e. research institutes, SMEs,
providers of software and technological solutions in the field of speech and language
technologies, technology developers).
External reviewers from the European Commission.
The main profile of participants included investigators, intelligence analysts and operational personnel
carrying out investigations in the field, providing their valuable points of view on the needs and
requirements of end-users. In addition, people active in the field of speaker recognition, such as
practitioners with a thorough experience of automatic systems, were present.
The second-day event attendees came from the following organisations:
Australian Federal Police
Home Office Digital, Data and Technology
2 Each INTERPOL member country maintains a National Central Bureau (NCB) linking national police with INTERPOL’s global
network of NCBs. The NCBs are staffed by national law enforcement officers and typically located as a division of the national
police agency or investigation service.
Page 6/23
Deliverable D9.4: Field Testing and End-User Training Final Report
Counter Terrorism Unit, Belgium Federal Police
Digital Security, EURECOM
Division of Identification and Forensic Science
European Commission ERC
ENFSI Expert Working Group Forensic Speech and Audio Analysis
Estonian Forensic Science Institute
Federal Police Limburg, Belgium
Forensic Bureau, International Security Agency, Poland
Forensic Science Centre of Lithuania
L3S Research center, Leibniz University of Hannover
Greek Police
Inter-ministerial control group (GIC), France
Financial Guard (GdF), Italy
Hungarian Institute for Forensic Sciences
Internal Security Forces, Lebanon
Institute for Advanced Technologies, University of Virginia
Ministry of the Interior of Croatia
National Centre for Biometric Studies Pty Ltd Australia
National Crime Agency UK
National Central Bureau Bissau
National Central Bureau India
National Central Bureau Rabat
NEC Japan
National Forensic Institute, Romania
French National Police
Phonexia, Czech Republic
Police Service of Northern Ireland
Federal Police of Brazil
Punjab Police, India
French Forensic Police Office
State Forensic Examination Committee of the Republic of Belarus
Swedish Security Service
Swiss Federal Criminal Police
Page 7/23
Deliverable D9.4: Field Testing and End-User Training Final Report
Given the multidisciplinary nature of the SIIP project, throughout its lifespan, the project benefited from
a global pool of experts in the field of speaker identification technology, comprising law enforcement,
private sector and academia. Certain attendees of the Field Test have already engaged in past SIIP events
and activities, by participating in the previous Filed Test organised by the Portuguese Criminal Police in
April 2017 or the Proof of Concept event held in June 2016 or by providing their feedback to the SIIP
questionnaire circulated at the project’s launch in July 2014.
4 DISCUSSIONS
This section covers the content of the discussions held during the second day meeting, the final Field
Test.
4.1 Key note speech
,
INTERPOL, inaugurated the event. As part of his
keynote speech,
highlighted the heavy work that SIIP has done to explore the operational,
technical, forensic, legal, societal and ethical perspectives related to the evolving field of speaker
identification. Furthermore,
reminded the audience of the increasingly changing landscape
shaped by on-going technological developments, under which LEAs worldwide have to operate. The
INTERPOL
stressed the potential operational value that the
combination of speaker identification with other biometric technologies such as fingerprints and facial
recognition has, and how it could potentially enhance investigative capabilities and facilitate checks
made against INTERPOL’s databases.
4.2 SIIP overall presentation
Following the opening remarks by
, the project coordinator,
from Verint,
provided an overview presentation of the SIIP project.
presented the project’s key aspects
and outlined the project’s milestones. The project coordinator reminded the audience of the SIIP
research objectives (technological SIIP system use cases, info sharing procedures and methodologies,
legal, ethical and societal acceptance) and the consortium expectations from the final Field Test
(demonstrate the developed system, disseminate the project results and receive feedback from experts
and end-users). The project coordinator concluded his intervention by inviting all the event guests and
participants to introduce themselves, mentioning their background and project engagement.
4.3 Demonstration
To display the SIIP system’s innovative capabilities and global info sharing mechanism, two LEA project
partners (the UK Metropolitan Police and the Portuguese Criminal Police) performed live demonstrations
of the system and its contribution to law enforcement work, while IDIAP presented the fusion
algorithms.
The Metropolitan Police field test scenario focused on the identification of an unknown hate speech
propagandist preaching in both English and Arabic on different Internet-supported channels. The
unidentified speaker was compared to a reference database of known preachers. The SIIP system is
capable of capturing speech files from a variety of sources, including open source internet search and
imported material.
The Portuguese Criminal Police executed a scenario involving a robbery investigation based on lawfully
intercepted mobile telephone conversations in Portuguese.
Page 8/23
Deliverable D9.4: Field Testing and End-User Training Final Report
Both scenarios used real data. As part of the demonstrations, each law enforcement authority
representative demonstrated the system’s operation step by step: creation of case management file,
creation of speaker of interest, data importation, segmentation and enrolment, creation of known
speaker model, processing of audio for search against reference database, sharing audio with SISC (SIIP
Sharing Center), analyses of output. The process was explained to the audience detailing the operation
of the overall SIIP eco-system. The demonstration described the three levels of the system, precisely
national police headquarters, centralized secured database at INTERPOL for analyzing and comparing
voice samples and the tactical level based on portable devices such as smartphone or tablet. Based on
this architecture, the SIIP eco-system entails an international info sharing mechanism of voiceprints and
metadata. The SIIP-SISC portal operates on four pillars consistent of core, collection, enrollment and
global. The system’s functionalities were covered, i.e. age, gender, language, accent detection, search
and filter database of reference speaker files. In both demonstrations, intelligence operators
incorporated the automated speaker recognition engine outputs (i.e. ranked list of potential matches
based on scores to confidence and likelihood ratio to scores) with other available relevant information.
Through these simulated live scenarios displaying the use of the SIIP portal based on real cases, the
audience witnessed the system’s innovative probabilistic, language independent capabilities based on
the fusion of a variety of Speaker Identification (SID) engines able to identify unknown speakers in social-
media/OSINT and lawfully intercepted channels. The tool can also be applied to recordings made directly
at a crime scene. These demonstrations confirmed that the developed solutions are feasible in a variety
of LEA investigation environments. Finally, the guests were shown the SIIP system’s secured and privacy-
preserving model built on privacy by design architecture.
4.4 Enrichment presentations
Following the SIIP system’s live demonstrations, a number of enrichment presentations were made by
different consortium partners to complement the Field Test event.
4.4.1 German Federal Police
from the German Federal Police (BKA) Language and Audio department (KT34) made a
presentation on the “SIIP Benefits for Investigation and Intelligence”.
gave an overview of
the state of forensic speaker recognition research and development (existing projects, applicable
international standards, norms and requirements). He then provided examples of the current application
of voice biometry (commercial speaker identification such as telebanking and forensic voice comparison
as part of prosecution and criminal courts activities).
described the benefits of voice
biometry for investigation and intelligence, highlighting the SIIP solution’s impact and potential
exploitation from a European perspective. The identified benefits included:
SIIP results confirm the potential applications of automatic SID using language-independent,
multi-source database applications for investigative, intelligence purposes;
SIIP has the potential to tackle transnational threats, terrorism, organized crime;
SIIP complies with the Internal Security Strategy (ISS) of cooperation between EU LEAs and
relevant third countries;
SIIP delivers the necessary infrastructure for secure cross-border exchange of speech, speaker
data.
Nevertheless,
cautioned that the human factor plays an important part within the SIIP expert
system, so it should be taken into consideration alongside the knowledge and training aspects.
Furthermore, he suggested that more testing with real data was required.
Page 9/23
Deliverable D9.4: Field Testing and End-User Training Final Report
4.4.2 International Biometric Group
The International Biometric Group made an intervention on “Biometric Liveness and Presentation Attack
Standards”.
started by defining the terms. Presentation attack refers to presentation to the
biometric data capture subsystem with the goal of interfering with the operation of the biometric system
in order to impersonate or obfuscate someone’s identity. Presentation attack detection (PAD) is an
automated determination of a presentation attack, and liveness detection is a PAD technique employed
by some biometric systems to detect presentation attacks. Turning on to the issue of standardization of
presentation attack detection, the presenter outlined some of the reasons behind the standardization
efforts (e.g. PAD susceptibility undermines the utility of biometric systems, to enable measure and
understanding of PAD subsystem performance, etc.) and listed the published standards (i.e. ISO/EC
30107-1:2016, ISO/EC 30107-3:2017). Further on, some of the key concepts addressed in the
international standards were explained (i.e. sensor/capture device, presentation attack instrument) and
the measuring of PAD performance explained (12 PAD performance metrics, measuring places, etc.). Mr
Thieme concluded his intervention by presenting some of the challenges and complications related to
PAD and its future, mentioning how resource-intensive testing was, the effect of rapidly changing state
of the art and the impact of machine learning on PAD development.
4.4.3 New Scotland Yard
from the National Digital Exploitation Service (NDES) within the Counter Terrorism Policing
at New Scotland Yard presented on the “Operational Potential of the SIIP system”. In line with NDES’s
vision to be a world leading service, it plans to make the best use of new technology and strives to
ensure that capabilities are developed and utilized efficiently. Within this context, the NDES operates the
Digital Seven services (for the Metropolitan Police London) with responsibility for Communications Data
Exploitation, Open Source Exploitation, Digital Media Exploitation, Lawful Interception, Technical
Innovation and Development, Digital Biometric Exploitation, Digital Operations. With the exponential rise
in media uploads, the Counter Terrorism Internet Referral Unit was formed in 2010. This unit is in charge
of identifying harmful online content and seeking to remove it by contacting the hosting companies.
highlighted the potential of the SIIP solution to respond to future challenges by offering the
global law enforcement environment a workable framework taking into consideration the ethical, legal
and societal aspects. The identified SIIP use cases and benefits included:
Lawful interception;
Identifying persons responsible for the upload of harmful media;
Identifying unknown third parties collected through covert or overt policing activity;
Collaborative working to maximise language resource;
Utilising the academic findings to assist in the creation of lawful and ethical frameworks.
4.4.4 Portuguese Criminal Police
The Portuguese Criminal Police made a presentation on the operational uses of SIIP.
,
within the Portuguese Criminal Police, started by presenting the national legislative context
(Internal Security Law n°58/2008 and the provision on control of communications, as well as the
restrictive data protection requirements) within which SIIP would have to operate.
defined
the operating monitoring system (e.g. all recordings in stereo), the tools used (Batvox Nuance software)
and listed the means from which data can be collected (e.g. lawfully intercepted calls, emergency calls,
internet support). On this background,
highlighted that SIIP could be used to perform fast 1 to
N speaker recognition and speech recognition, whilst tactical SIIP can be used on the field when the voice
of the target is known. Additionally the evidence processing and analysis use offered by SIIP was
detailed, such as the creation of a case file in SIIP based on a request received from investigator,
provision of comprehensive reports on audio tests, upload of voice prints, queries within SISC and local
Page 10/23
Deliverable D9.4: Field Testing and End-User Training Final Report
databases (pending legislative revision). Another example of SIIP’s operational use is the support to real-
time field operations and quick tests given the system’s simple interface and its ability to interconnect
with field devices, and to accept data from various audio sources. Concerning judicial admissibility,
highlighted that according to Portuguese legislation, forensics voice examination does not qualify
as evidence but would contribute to the judicial decision. Finally,
highlighted that the results
obtained from 1-to-N comparisons performed by SIIP are pre-forensic, meaning that to be presented in
court, it will always be necessary to ensure the suspect reference sample for forensic comparison and
identification 1-to-1 (as in fingerprints, DNA and any other biometric method) and does not qualify as
evidence but would contribute to the judicial decision. Mr Freitas outlined some of the encountered
technical difficulties such as the inherent features of old recording systems, the short duration of useful
audios and the degradation of the original quality of audio files during conversion to requested
parameters.
4.4.5 Netherlands Forensics Institute
Finally, a presentation by
from the
Netherlands Forensics Institute, was cancelled due to travel impediments.
input focused
on providing some recommendations to the SIIP system once it becomes operational, such as ensuring
support by technical staff and forensic speech experts, and database maintenance (e.g. back-up, update).
5 END-USER TRAINING
The original objective of the training session was to organise a set of workshops dedicated to the training
of potential SIIP end-users (technical personnel, intelligence analysts and operational personnel carrying
out investigations) on the operation, features and capabilities of the developed system. Given the
composition of the Field Test participants as described in Section 2, comprising law enforcement
representatives from a diverse background, one can conclude that the targeted audience was
successfully gathered. However due to the large turn-out, exceeding the expectations at the time of
drafting the SIIP DoW, the individual hands-on training session had to be converted to a lecture type
public training held in INTERPOL’s main conference hall. To facilitate the training session, a hard copy of
the End-User Manual was distributed to participants, enabling them to simultaneously follow the
explanations and refer to the Manual when necessary.
Singular Logic conducted the training session, navigating the audience throughout the SIIP system and
explaining its features and capabilities as demonstrated during the field environment scenarios earlier in
the day by the Metropolitan Police and the Portuguese Criminal Police. In addition, the trainer referred
to the End-User Manual and SIIP portal updates. Singular Logic expanded on the content of the End-User
Manual, which includes definition of main use cases and of execution scenarios implementing those use
cases. The end-user feedback provided throughout the project’s lifespan contributed to the revision of
the preliminary Manual and compiled training material. As a result, an extra chapter was added
explaining media upload and import, flow diagrams were included in each chapter, further information
was provided to ensure greater clarity. In addition to Manual revision, end-users’ input was integrated as
part of the technical work packages to enhance the developed tools and individual capabilities, to modify
the user interface portals (WP7) and to refine the different evaluation criteria (WP9). Among the portal
updates was mentioned the addition of a new use case, and the UI/UX enhancements e.g. display of
audio file’s name under speaker’s profile, visibility of engine results in case of no speaker match. To
ensure an interactive training session, the floor was open to the audience which to intervene and raise
questions or request additional information as to the system’s performance.
Page 11/23
Deliverable D9.4: Field Testing and End-User Training Final Report
6 DISSEMINATION
Various means have been employed to inform the stakeholders about the organization of the event and
to ensure the widest dissemination of the Field Test results. Prior to the event, a notice was published on
the INTERPOL official website3 informing about the purpose of the Field Test, the sought expertise of
participants and including a link to the SIIP website, where registration for the event was possible.
Figure 1. Print screen of the informative note about the SIIP Field Test on INTEROL’s website
3
https://www.interpol.int/News-and-media/Events/2017/Speaker-Identification-Integrated-Project-SIIP/Speaker-Identification-
Integrated-Project-SIIP-Field-test-2
Page 12/23
Deliverable D9.4: Field Testing and End-User Training Final Report
The project’s website also displayed a Field Test announcement which included a description of the
event,
invitation,
event
agenda,
SIIP
factsheet
and
SIIP
brochure.4
Figure 2. Print screen of the registration page on the SIIP website
A video promoting the SIIP project and explaining how the system works was published on the INTERPOL
official website5 and the SIIP website.6 The video provides an overview of the project, the consortium
members and explains how voice recordings are collected and processed and how this could serve to
identify criminals.
4
5 https://www.interpol.int/News-and-media/Videos/(video_id)/35605
6
Page 13/23
Deliverable D9.4: Field Testing and End-User Training Final Report
Figure 3 & 4. Print screen of the SIIP explanatory movie available on the INTERPOL and SIIP websites
INTERPOL updated its SIIP factsheet (available in French, Spanish, English and Arabic) in order to reflect
the latest project developments, such as the outcome of the Proof of Concept event held at the
Carabinieri School in Rome in June 2016 and the contribution of the first Field Test organized in Lisbon in
March 2017. This latest version of the SIIP factsheet, available in Annex, also announced the final Field
Test to be hosted by INTERPOL.
The page dedicated to the SIIP project on the INTERPOL website was also updated accordingly.7
Figure 5. Print screen of the SIIP page on the INTERPOL website
The INTERPOL Communications departments provided live event support during the SIIP Field Test,
ensuring filming, photography and post-production, as well as web support with press release drafting
and publication. Some of the pictures were published afterwards on the SIIP website.8
7 https://www.interpol.int/About-INTERPOL/Legal-materials/ICT-Law-Projects/SIIP-Project
8
Page 14/23
Deliverable D9.4: Field Testing and End-User Training Final Report
Figure 6. Picture taken during the Field Test event
The video material prepared includes a flash video report9 published on YouTube under the INTERPOL
channel, as well as videos of the different individual presentations.
Figure 7 & 8. Print screen of the video material available on YouTube and INTERPOL website
In addition, a video of the key note speech made by
INTERPOL
was published on the INTERPOL official website. 10
Following the event, a press release was published on the INTERPOL internal website highlighting the
event outcomes to a worldwide audience of 192 member countries.
9 https://www.youtube.com/watch?v=i7jEEyYiCjk&feature=youtu.be
10 https://www.interpol.int/News-and-media/Videos/(video_id)/38492
Page 15/23
Deliverable D9.4: Field Testing and End-User Training Final Report
Figure 9. Print screen of the press release published on the INTERPOL internal website
7 CONCLUSION
This report covered the SIIP Field Test and End-User Training held at the INTERPOL General Secretariat in
November 2017. This event represented the second and final field test organised to run the SIIP system
in order to demonstrate its performance and capabilities before the project’s closure. The event also
provided an opportunity to familiarise a large audience of stakeholders (investigators, intelligence
analysts, technical staff and operational personnel carrying out investigations in the field) with the
system’s operation and capabilities, and to distribute associate training material, i.e. End-User Manual.
The event achieved the objectives set in the DoW referred to in
Box1, and responded to the participants’
expectations based on the positive feedback received from the consortium partners and the other
attendees.
After close interaction and validation of SIIP system by dozens of LEAs from all over the world during the
proof of concept and the two field test events, it appears that there is significant demand and need for
SIIP system among LEAs as it may expedite significantly the police work and assist in delivering a much
more efficient investigatory process in the area of voice recognition, supplemented by the human factor
as highlighted throughout the project. The role of advanced speaker identification in tackling global crime
is constantly increasing and SIIP provided a unique platform for testing and deployment of cutting-edge
technologies for potential use by law enforcement around the world. Furthermore, the results derived
from the SIIP assessment and validation actions will feed into the recommendations regarding the
extensibility of the SIIP tool to the security field.
Page 16/23
Deliverable D9.4: Field Testing and End-User Training Final Report
ANNEXES11
ANNEX 1: Group photo
11 Double click on the annexes to open the full PDF file
Page 17/23
Deliverable D9.4: Field Testing and End-User Training Final Report
ANNEX 2: Event agenda
Page 18/23
Deliverable D9.4: Field Testing and End-User Training Final Report
ANNEX 3: Template invitation letter
Page 19/23
Deliverable D9.4: Field Testing and End-User Training Final Report
ANNEX 4: SIIP updated factsheet
A) English version
Page 20/23
Deliverable D9.4: Field Testing and End-User Training Final Report
B) French version
Page 21/23
Deliverable D9.4: Field Testing and End-User Training Final Report
C) Spanish version
Page 22/23
Deliverable D9.4: Field Testing and End-User Training Final Report
D) Arabic version
Page 23/23