Check this out
http://news.teamxbox.com/xbox/19960/Rumor-New-Xbox-Console-in-2010-We-Tell-You-the-Truth/so im looking for ibm video-based motion sensing and the first thing i get?
Digital Video Surveillance
http://www-935.ibm.com/services/us/gts/pdf/sp_wp_digital-video-surveillance.pdf( and something is messing with my browser making links go elsewhere.. grr)
anyhow, Digital Video Surveillance:
enhancing physical security with
analytic capabilities
by IBM global services 2008
Today’s Surveillance Challenges
In today’s environment, virtually every municipality, agency, educational
institution, mass transportation center, financial institution, utility plant and
medical center must plan for threats and protect the security of its property,
employees, customers, citizens and IT infrastructure. Additionally, businesses
in every sector face challenges in protecting their customers, employees, and
assets while working to reduce operating costs, improve productivity and
increase profit as well as customer satisfaction. Examples of security risks
and business issues that may be managed more effectively using surveillance
methods include:
• Public Safety/Security: Increased threats have caused many government
agencies to deploy surveillance cameras and sensors, providing situational
awareness around critical facilities. School campuses must deal with pro-
tecting entry and exit points, maintaining IT network security, preventing
vandalism and avoiding authorization issues.
• Airports/Seaports/Railways: Mass transit businesses and agencies
must protect passengers, staff, and physical assets from terrorist threats
and security breaches, and adhere to regulatory requirements.
• Retail Stores: The retail industry monitors establishments to reduce
fraud, theft, and administrative errors. Retail stores also use video and
analytic information to determine the effectiveness of promotional
displays and count people in various areas to optimize store layouts as
well as sale effectiveness.
• Financial Institutions: Many banks have 24—hour human surveillance
requirements for inside operations and Automated Teller Machines (ATMs).
Surveillance and analytics are being used to reduce threats of robbery as
well as fraud. Many banks are consolidating security controls across bank
branches by monitoring video, voice and transactional information from a
central command and control center.
An Evolution of Surveillance
Organizations have used surveillance for decades as a deterrent to criminal
activities such as theft, fraud, and violence. In the last ten years, surveillance
technology has been developed that not only helps organizations detect and
respond to threats sooner, but also helps them focus on improving business
operations. The three generations of surveillance are often described as:
• Analog
• Digital
• Smart or Intelligent
We will explore each of these in greater detail to help you better understand
how we arrived at today’s environment and where we plan to go in the future.
Analog Video Surveillance
Video surveillance has typically involved the placement of analog video
cameras in sensitive or strategic areas of a particular business, coupled with
closed-circuit television (CCTV) for live monitoring. This serves not only as
a deterrent to crime, but also to record the movement of people and property.
Mobile methods of video surveillance, such as mounting cameras in patrol
cars, buses and trains are also often utilized to record events.
The use of analog video cameras results in the creation of hundreds of video-
tapes that then must be viewed by security guards. The cost of employing
security personnel to monitor hundreds of cameras, in addition to storing a
high volume of videotapes can be prohibitive. Additionally, videotapes can
have poor image quality and deteriorate over time.
More importantly, studies have shown that a person assigned to sit in front of
a video monitor for several hours a day and watch for particular events is an
ineffective security system. Tests have demonstrated that after only 20 minutes
of watching and evaluating monitor screens, the attention of most individuals
has degenerated to well below acceptable levels.1 Monitoring video screens is
both boring and mesmerizing. Furthermore, manual searches of tapes can take
too long to provide vital information needed to assist in investigations.
Also, video can often only be viewed from a single end point that is not shared.
This limits the ability to distribute information across an enterprise, which
could help minimize company-wide threats and alerts. Finally, analog video
systems cannot extract business intelligence from security data.
Digital Video Surveillance
Today, video surveillance remains as vital as ever, but it assumes a new role.
The emergence of digital video, IP video cameras, networked video recorders,
web video, consumer cameras and video-based intelligence is opening up a
wide range of applications providing enhanced functionality and business
value to organizations.
Digital video surveillance (DVS) enables clients to establish effective risk
management strategies that will help them manage and safeguard business
information and technology assets, anticipate vulnerabilities and risk, and
maintain timely access to information.
Many organizations have piecemeal solutions, and are challenged by
having multiple systems that do not communicate with each other. Often,
the separation of IT and physical security does not allow organizations to
take advantage of existing IT infrastructures and applications, such as
identification (ID) management and transactional systems that may already
be in place. Operating totally separate, disparate systems for IT and physical
security is not only less effective, but also more labor intensive and costly.
Figure 1. Digital video surveillance drives intelligence through integration, which can
enhance your ability to respond.
Migrating to a DVS solution will help address some of the limitations of a tape-
based analog system. DVS can help organizations achieve better returns on
their security investments by:
• Enabling real-time detection and potential prevention of security incidents
through enhanced intelligence gathering
• Using event-based viewing for investigative purposes, eliminating the need
to chronologically review videotapes
• Reducing the need to monitor video cameras and change tapes
• Increasing product security by deterring potential shoplifters and
monitoring staff
• Providing evidence against fraudulent claims
• Increasing indoor and parking lot security
Smart Surveillance
Smart surveillance, intelligent video surveillance, video analytics, intelligent
video and intelligent analytics are typical names used to describe the concept
of applying automated signal analysis and pattern recognition to video cameras
and sensors, with the goal of automatically extracting “usable information”
from video and sensor streams.IBM Smart Surveillance Solution (SSS) helps optimize security by integrating
hardware, software and services within an organization, thereby enabling the
convergence of physical and IT security. An integral part of SSS is a software
component developed by IBM Research known as IBM Smart Surveillance
Analytics (SSA), which provides capabilities that enable real-time decision
making and post-event correlation of people and activities.
Usage Scenarios
IBM Smart Surveillance Analytics has many unique features to help clients
manage security issues and prevent problems before they occur, such as:
• Open framework—A comprehensive security and surveillance plan may
involve multiple modalities of events captured from various video analytic
technologies, non-video sensors and event systems like TLOG in the retail
environment. SSA has been designed with an open framework that enables
event-based surveillance and can make the integration of events simple
and easy.
• Behavior factory—Many vendors provide a set of behaviors, such as
“large fast vehicle” and “stopped vehicle.” These behaviors have been
designed with a limited customer set in mind. SSA’s ability to alert based
on “database indexed metadata” of all events occurring across a series of
camera feeds allows the user to customize behaviors to their environment
through an easy-to-use interface.
• Attribute search—The intelligent video industry has approached surveil-
lance based on a limited set of known threat models; hence, the emphasis
on “tripwires and abandoned objects” and very limited functionality
to support investigation of “unknown threats.” SSA, through its unique
and patent pending metadata search, supports a wide range of queries on
events that may or may not have been previously defined as alerts. This is
possible because SSA is capturing metadata on event activity, not just on
pre-defined alerts.
AND THAT INCLUDES
• Entertainment solutions—The capability of SSA to track people can be
used at sporting events to generate enhanced statistics, visualizations
and interactive gaming. Clients that may be interested in this capability
include casinos, sporting leagues and television stations.
IBM Smart Surveillance Solution Architecture
As stated above, the integral software component of SSS is IBM Smart
Surveillance Analytics. The analytic framework of SSA is comprised of
two core components: Middleware for Large Scale Surveillance (MILS)
and Smart Surveillance Engine (SSE). They will be discussed later in
this paper.
SSA provides the unique capability to carry out efficient data analysis of
video sequences, either in real-time or recorded video. Based on open
standards-based middleware, the open standards-based software platform
is designed to allow monitoring and analysis of real-world events via sensors
(like video cameras, radar or audio inputs).
All SSA functionality is Web-based, allowing virtually “anytime, any-
where” access to both real-time and historical event data from the system.
Figure 2 shows the high-level conceptual architecture of IBM Smart
Surveillance Solution. It illustrates how Smart Surveillance Analytics
integrates with existing video cameras and capture systems to provide:
• Video/sensor analytics capabilities
• A framework for integrating event information from multiple
related sources
• A framework for building client-specific solutions drawing upon the video
and sensor events and integrating these into the client’s business process
Figure 2. IBM Smart Surveillance Solution—a conceptual architecture
SSA provides the following types of functions to the end user:
• Real-time alerts: Users can specify “alert definitions” that include multiple
conditions from a single camera/sensor or across multiple cameras and
sensors. SSA uses its analytics capabilities to evaluate events occurring
in relevant sensors against the alert definition. Each time the “alert
definition” is triggered, SSA can provide the user with prompt notification
of the event.
• User-driven queries: Users (both human and applications) can use SSA to
perform content-based queries against event metadata that is archived by
SSA. For example, SSA can retrieve all events from a camera where “a red
car” was driving across the parking lot.
Figure 3. SSA Functionality
• SSA Functionality: There are several types of video analysis technologies
that are part of SSA. Typically, each of these analytics involves sophisti-
cated algorithms that process the video/sensor signals to extract
information and structure the information to support the real-time alert
and search functionalities of SSA. The video analysis technologies are:
– Behavior analysis: These analytics are intended to analyze the move-
ment of objects within the field of view of a camera. This is based on the
ability to detect and track multiple moving objects across the camera,
classify these objects, and extract various object attributes like color,
shape and size. The extracted information is used to provide a variety of
alerts while recording information from all events (for example, motion
detection, tripwire, abandoned object) and search functions
(for example, find red cars).
– License plate recognition (LPR): This analysis capability is tailored to
detect the presence of text within a given video frame, and apply optical
character recognition technology to extract the license plate number.
LPR needs to be customized to the character set (for example, English,
Arabic), style, format and appearance of the license plate, which varies
significantly across geographies. In order to correctly operate, LPR
requires a minimum resolution across the license plate and adequate
illumination and viewing angle.
– Face detection: This analysis capability is designed to automatically
detect images of human faces from the video. The face detection
capability creates an index in the video and marks the time at which
the face was present in the video. The system generates a key frame to
represent the face, thus producing a catalog of faces for all the people
who appeared within a camera field of view (approaching the camera.)
– Event integration: This capability allows the integration of events
from the analysis of other sensors (like automatic door sensors, HVAC
sensors, audio) with event streams from other IT systems (like point-of-
sale, telephone call logs). Once integrated, the event information can be
cross-correlated to video-based events like behavior analysis, LPR
and face detection.
Figure 4. SSA Software Architecture
SSA has the following core components:
1. Smart Surveillance Engine (SSE): The Smart Surveillance Engine (SSE)
is a C++ framework for capturing events that are observed by sensors such
as cameras. Every physical camera in your environment is assigned to an
analytic engine running on an SSE server. One SSE Server can handle
multiple cameras. In general, SSE is designed to process streams of video
in real-time, automatically extracting event (activity in the camera’s field
of view) metadata and evaluating user-defined alerts. The specific user
functionality associated with each camera is based on the profile which is
configured for use by the analytic engine associated with the camera. The
following profile types are available:
• Behavior analysis (Outdoor Far-field, Outdoor Near-field, Indoor
Tracking)
• Face detection (Face Tracking, Sensitive Face Tracking)
• License plate analysis (via integrated IBM Business Partner technology)
The information extracted by SSE from the camera’s field of view is used to
classify objects according to the profile type, providing metadata on object
type, object size, object speed, etc.
SSE alerts are conditions which have been specified by the user as being
of interest. SSE supports both basic video alerts and compound metadata-
based alerts. Currently, we support the following basic video alerts:
• Motion detection-detects motion within a specified region of view
• Tripwire-detects directional crossing of user-defined tripwire
• Region-detects certain specified behavior within a specified zone,
such as entering, leaving, starting and stopping
• Abandoned object-detects when an object has been left behind
• Object removal-detects when an object has been taken away
• Directional movement-detects when objects are moving in a user-
specified direction
• Camera move/blind-detects changes in camera state such as movement
or obstruction
• Camera movement stop-detects when a pan-tilt-zoom (PTZ) camera
stops moving
2. Middleware for Large Scale Surveillance (MILS): Each installation of
the IBM Smart Surveillance Solution, which includes SSA, has a MILS
Server. MILS is a J2EE framework application built around IBM’s DB2®,
WebSphere® application server and MQ platforms. In addition to metadata
management, MILS provides system management, user management
and various extensibility services, including a web services application
programming interface (API).
MILS can help provide consolidated backend data management capabilities
and store metadata that describes key activities discovered while ingesting
video data. It can also create and manage a full index of the ingested video
data. This index has a full set of event attributes that can be searched to
support forensic analysis. The index attributes can also be used to define
composite metadata-based alerts by combining the metadata in various
ways to define complex behavioral patterns.
MILS operates on top of a software middleware stack, either IBM Web-
Sphere Remote Server (WRS) or IBM Central Site Server. Both provide
a middleware platform with a J2EE application server called WebSphere
Application Server Network Deployment, integrated with:
– WebSphere MQ for an assured message delivery component
– DB2 Workgroup Server Edition as a relational database
management system
3. Applications: These are mainly web applications (HTML, Java, JSP, applets,
Javascript) which use the web services enabled by MILS to provide the
functionality needed by the user.
IBM Smart Surveillance Analytics can also allow administrators to add
new metadata schemas to the system, thus enabling new analytic engines
to send sensor/event metadata. The metadata from all analytic engines
can be consolidated allowing users to search across modalities. These
advanced indexing capabilities offer a unique and powerful differentiator
from virtually all other available surveillance solutions.
Return on Investment (ROI) by Industry
Implementing IBM Smart Surveillance Solution, which includes Smart
Surveillance Analytics, offers many benefits, including the potential to
increase return on investment (ROI). ROI successes fit into three cate-
gories: managing risk, growing the bottom line and growing the top line.
ROI highlights in various industries include the following scenarios:
1. Retail
In today’s retail environment, product shrink dramatically affects the
top and bottom line. Globally, on average, 1% to 3% of all retail sales are
affected by product shrink, due to conditions including crime, employee
fraud, and damaged goods.2 This results in a significant impact on retail
margins, especially for those businesses running on a 1% to 3% margin.
IBM Smart Surveillance Solution can serve as a loss prevention tool as
well as a source of retail intelligence data. It can provide video techno-
logy to help manage profit and loss at the cash registers, under the
cash registers and throughout the store. Retailers can implement SSS
to determine promotion effectiveness, cashier monitoring and people
counting. Grocers can use the technology to help reduce Bottom of Basket
(BOB) losses. One grocer has reduced BOB by more than 80%, integrating
IBM’s optical recognition and an IBM Business Partner Point-of-Sale
(POS) system.3
all for profit.. but what do you expect when you find this, searching for the next xbox
