| |
Overview
Humans have a natural ability to recognize
objects. Once we've seen an object, we can recognize that same object again,
whether it is in the same or a different environment. But there are a growing number of tasks – from security to
manufacturing – where it is more efficient,
safe or cost-effective for the process of object recognition to be automated.
In these cases, a computer with object recognition capabilities is used in
conjunction with a camera or other visual input device.
The process of automating real-world object
recognition requires very specific algorithms. These algorithms enable the
computer to "recognize" an object that is presented to it by
comparing it to a list of objects that have been stored in the computer's memory
or a database. In order to be more like the human visual recognition process,
this special type of algorithm must be able to recognize objects at different
scales, rotations, poses and with variations in lighting. Oftentimes such
technology applications require real-time processing, so the speed of the
object recognition algorithm is also very important.
The technology
Sentisight object
recognition technology enables the learning of objects and searching for
learned objects in images from almost any camera, webcam, still picture or live
video. The technology has these features:
• Accurate. The SentiSight algorithm provides simultaneous multiple 2D and 3D object detection and recognition, and is able to find out:
° whether a particular object is presented in a scene;
° where the object is located in the scene;
° how many instances of the object are there in the scene.
The algorithm is also able to evaluate the region an object occupies in a scene, providing additional information about the size, orientation and scale of the recognized object.
• Robust. SentiSight features high recognition quality and is tolerant to object scale, rotation and pose. The algorithm is able to compare and identify pictures even when the perspective has changed.
• Universal. The SentiSight algorithm is
designed to be as universal as possible. It can support web cameras,
surveillance cameras and can input images from the picture. It is tolerant to
object scale, rotation, pose etc.
• Fast. SentiSight can process video streams
in real time, so it can be used for real-time applications.
• Webcam capable. Though high quality cameras
will provide better recognition quality, a simple webcam is enough for
SentiSight operation.
SentiSight 2.0 algorithm demo application
is available for downloading.
Read more about the
technology.
The SDK
SentiSight 1.1 SDK
is based on SentiSight object recognition technology and is intended for
developers who want to use computer vision-based object recognition in their
applications. SentiSight 1.1 SDK enables fully automatic and manual object
learning as well as simultaneous multiple object detection and recognition in
an easy, yet versatile, way.
SentiSight 1.1 SDK supports Microsoft
Windows operating system. You can download trial
version of SentiSight 1.1 SDK to try it for creating your system. Read more¡¦
Why SentiSight?
Neurotechnology's SentiSight technology
is intended for developers who want to use computer vision-based object
recognition in their applications. SentiSight enables the learning of objects
and searching for learned objects in images from almost any camera, webcam,
still picture or live video.
• Universal. The SentiSight algorithm is
designed to be as universal as possible. It can support web cameras,
surveillance cameras and can input images from the picture. It is tolerant to
object scale, rotation, pose etc. Some of the potential applications for
SentiSight technology include:
° Search engines that recognize objects in
picture files (either local or on the Web);
° Security systems;
° Parts recognition in production lines;
° Robot vision;
° Road signs recognition;
° Machine vision.
• Fast. SentiSight can process video streams
in real time, so it can be used for real-time applications.
• Webcam capable. Though high quality cameras
will provide better recognition quality, a simple webcam is enough for
SentiSight operation.
• Flexible licensing model and competitive
price.
Algorithm
The SentiSight 1.1 object recognition
algorithm implements advanced visual-based object learning and recognition.
|
1. Object learning
In order to recognize an object in an
image, the appearance of an object must first be memorized. In the learning
phase, SentiSight algorithms extract specific object features from a video
stream or single image and save them into what is known as a model template.
In many cases there is more information in
a video or single image than just the object you want SentiSight to learn, like
a background, other objects in the room or a hand holding the object. For this
reason, certain steps should be taken during the "learning" process
to provide information about the exact location of the desired object in the
image. This should be done with a mask of the object in the image. A mask
explicitly specifies which pixels of the image belong to the object and which
ones are part of the background. Thus, only object-specific information will be
included into the model template.
If there is no way to provide a mask for
the image, SentiSight can still learn the object. However the other background
elements would be learned together with the object. This can affect the ability
of the algorithm to recognize the unique qualities of the object and may result
in the object being misclassified with other objects that have the same
background.
However, for a lightweight movable object
SentiSight does provide a fully automatic learning procedure. To learn a
lightweight movable object in the SentiSight 1.1 SDK, the user should do
following steps:
1. Choose a static background (preferably
smooth) and direct the camera to it.
2. Choose a holder – an object that will be used to hold and move the learned object. A
user's hand can be the "holder".
3. This "holder" should be presented
it to the camera first, in various poses and configurations (if it is not rigid
object) so that it can be learned by SentiSight.
4. After the holder has been learned,
SentiSight is ready to learn the object itself, by having the holder rotate and
move the object closer and further from the camera.
5. Input the learned object name (ID) into the
system.
The process is somewhat different for
objects that cannot be moved or if only images of the objects are available. In
these cases, one should set the mask of the object manually or do not provide a
mask at all.
|
Background
learning
Click to zoom
|
|
"Holder"
learning
Click to zoom
|
|
Object learning
Click to zoom
|
| 2. Recognition
For the recognition of the object, the
camera is directed to the scene where the learned object is presented or may
appear. No other action is required. When the object appears in the vision
field, it is recognized by SentiSight, which outputs the object's name (ID) and
coordinates.
The SentiSight algorithm is tolerant to
large variations of object scale, object rotation and translation. When an
object is learned, the algorithm creates a model with possible views from
different sides, in different 3D poses and in different lighting conditions.
SentiSight's object recognition is
comparably fast – around 10 frames per second
for a single object model (320 X 240 resolution). However for tasks when an
even faster response is needed, the SentiSight 1.1 library has a tracking mode
that enables tracking speeds up to 20 frames per second. Tracking is
initialized if an object is recognized and located, then tracks the object
until it changes somewhat in appearance, at which point tracking is
reinitialized by recognition. The tracking feature is sensitive to complex
backgrounds, and tracking is more difficult with homogenous objects
All performance evaluations were performed
using a PC with 2.4 GHz Intel Core2 Duo CPU
|
Object
recognition
Click to zoom
|
Reliability Tests and Technical
Specifications
SentiSight 1.1 was tested with object
images from many cameras. At 0.1% False Acceptance Rate (FAR), the recognition
rate is from 70% to more than 99% depending on object structural appearance,
transparency, etc. For objects with well defined intenal structure, the
recognition rate is 98% - 99% at 0.1% FAR.
|
SentiSight 1.1 algorithm technical specifications
|
|
Recommended image size for real time operation on modern processor
|
320 x 240 pixels
|
|
Static Background Extraction / Object mask separation
(320 x 240 image size)
|
20 frames/sec
|
|
Learning: Processing of single objects' frame
(320 x 240 image size)
|
0.05 sec.
|
|
Learning: Generalization time
(for 100 frames of object)
|
6 sec.
|
|
Recognition speed from image frame for single object model
(including processing of the image)
|
~ 10 frames/sec.
|
| Recognition speed from image model for single object model
(excluding processing of the image)
|
~ 20 models/sec.
|
All Performance evaluations were performed using a PC with 2.4 GHz Intel Core 2 Duo CPU
System Requirements
• PC with 1GHz processor supporting SSE2
technology;
• 256 Mb of RAM.
Algorithm Demo
There are 2 SentiSight 1.1 demo
applications for Microsoft Windows 2000/XP/2003/Vista
that can be downloaded
for evaluation of the SentiSight 1.1 vision based object recognition algorithm.
| • SentiSight 1.1 full demo application learns
and recognizes objects from almost any camera or webcam, image and video files.
You may need to view SentiSight
Demo Usage Guide for more convenient evaluation of SentiSight technology.
• SentiSight 1.1 simple demo application is
intended for fast and simple evaluation of SentiSight 1.1 technology. The
application has wizard-based user interface and uses only webcam for object
learning and recognition.
|
|
Click to zoom
|
Internet connection is not required to run
both demo applications.
SentiSight 1.1 SDK trial is also available
for downloading.
SDK Overview
SentiSight SDK is based on the SentiSight object
recognition technology and is intended for developers who want to use computer
vision-based object recognition in their applications. SentiSight enables
learning of objects and searching for learned objects in images or video and
allows rapid development of applications using functions from the SentiSight
library. SentiSight SDK enables manual and fully automatic object learning as
well as simultaneous multiple object detection and recognition in an easy, yet
versatile, way.
SentiSight can be easily integrated into a
customer's system. The developer has complete control over SDK data input and
output; therefore SDK functions can be used in connection with most cameras
(including webcams), with any database and with any user interface.
SentiSight is available as SentiSight 1.1
SDK. The SDK includes Camera Manager Library for Microsoft Windows that allows
simultaneous capture from multiple cameras and a library to aid in handling
video files.
SentiSight 1.1 SDK
SentiSight 1.1 SDK allows the development
of computer vision-based object recognition systems for Microsoft Windows or Linux
platforms. The kit includes programming tutorials in several programming
languages.
SentiSight 1.1 SDK distribution package
contains:
• One SentiSight 1.1 installation license;
• SentiSight 1.1 learning and recognition
algorithm;
• C/C++ programming tutorials;
• C# programming tutorials;
• SentiSight 1.1 SDK documentation.
System requirements for SentiSight 1.1 SDK:
• PC with processor that supports SSE2*.
These processor families include:
° AMD K8-based CPUs (Athlon 64, Sempron,
Turion 64, etc.);
° Intel Pentium4 family (Pentium4, Xeon,
Celeron, etc.);
° Intel Pentium M and Celeron M;
° Intel Core- family (Core Duo, Core Solo,
etc.);
° Intel Core 2- family (Core 2 Duo, Core 2
Quad, etc.);
° Transmeta Efficeon;
° Via C7.
• Optional camera or webcam.
• Microsoft Windows specific:
° Microsoft Windows 2000/XP/2003/Vista.
° Microsoft DirectX 9.0 or later.
• Linux specific:
° Linux (based on glibc 2.5 or newer).
° Video4linux.
* Processors that do NOT support SSE2 cannot
run the SentiSight 1.1 algorithm. These include: AMD CPUs prior to Athlon 64
and Sempron, Intel CPUs prior to Pentium 4, Via C3, Transmeta Crusoe.
SDK Trial
Neurotechnology offers SentiSight 1.1 SDK
on a 30 day trial that allows developers to explore the SDK's possibilities and
to try it in real environments and real applications. Constant Internet
connection is required during evaluation. SentiSight 1.1 SDK trial is available
for downloading.
SentiSight 1.1 algorithm demo applications
for Microsoft Windows are also available for downloading.
Licensing SentiSight
To develop a product based on SentiSight
1.1 technology, an integrator should obtain SentiSight 1.1 SDK. Integrators can
develop only an end-user product using SentiSight 1.1 SDK and sell/install the
product to their own customers*.
A license for SentiSight 1.1 is required
for each running instance of SentiSight-based application. The following
license types are available:
• Single
computer license.
• Enterprise
license.
SentiSight 1.1 SDK includes one SentiSight
1.1 installation license.
SentiSight 1.1 SDK customers can obtain
additional SentiSight 1.1 licenses for their product installation or
development at any time. Prices for additional SentiSight 1.1 installation
licenses can be found here.
Please also refer to SentiSight SDK
Software License Agreement for all licensing terms and conditions.
Single
computer license
A single computer license allows to install
and run a SentiSight 1.1 installation on one computer. License will not be lost
if computer will be reinstalled.
The following license management options
are available:
• license activation online by communicating
with Neurotechnology server;
• license activation by email;
• license activation using volume
license manager;
• license management using volume
license manager on LAN or Internet.
SentiSight
1.1 enterprise license
SentiSight 1.1 enterprise license allows an
unlimited use of SentiSight 1.1 engine in the end-user products in the certain
territory, market segment or project. These limitations would be included in
the licensing agreement.
Each SentiSight 1.1 installation requires
activation. SentiSight 1.1 enterprise license customers receive software tools
that allow to activate an unlimited number of SentiSight 1.1 installations
without any communication with Neurotechnology.
For more information please contact us.
Volume license
manager
Volume license manager is used on site by
integrators or end users to manage obtained licenses for SentiSight. It
consists of license management software and a dongle, which are used to store
the number of obtained licenses. An integrator or an end-user can use the
volume license manager in the following ways:
• Activating the single computer licenses. An
installation license for a SentiSight 1.1 will be activated for using on a
particular computer. The license quantity in the license manager will be
decreased by the amount of activated licenses.
• Managing the single computer licenses on
LAN or Internet. The license manager allows to manage installation licenses for
SentiSight 1.1 across the computers on LAN or Internet. The number of managed
licenses is limited by the number of licenses in the license manager. No
license activation is needed and the license quantity is not decreased. Once
issued, the license is assigned to certain computer on the network.
• Using a license manager as a dongle. The
volume license manager containing at least one license for a SentiSight 1.1 can
be used as a dongle that allows to run SentiSight 1.1 installation on a
particular computer.
Additional SentiSight 1.1 installation licenses
for the license manager can be purchased anytime. Neurotechnology will
generate a special update code and send it to you. Then you will just have to
enter the code to the license manager to add these purchased licenses.
* If the integrator wants to develop and
sell a SentiSight 1.1 based development tool (with API, programming
possibilities, programming samples, etc.), he/she will need a Neurotechnology
permission and shall sign a special VAR agreement.
Download
• SentiSight 2.0 Algorithm Demo (for MS Windows)
• SentiSight 2.0 Algorithm Demo (for Linux)
• SentiSight 2.0 SDK Trial
• Data files for SentiSight Demo and SentiSight SDK Trial
Pricing
|
SentiSight 1.1 SDK (licensing model)
|
|
SentiSight 1.1 SDK
|
¢æ990.00
|
|
SentiSight 1.1 installation licenses (prices per license)
|
|
Quantity
|
Price
|
|
1
|
¢æ99.00
|
|
2-4
|
¢æ91.00
|
|
5-9
|
¢æ82.00
|
|
10-19
|
¢æ73.00
|
|
20-49
|
¢æ64.00
|
|
50-99
|
¢æ55.00
|
|
100-199
|
¢æ47.00
|
|
200-499
|
¢æ38.00
|
|
500-999
|
¢æ29.00
|
|
1000-1999
|
¢æ22.00
|
|
2000-3999
|
¢æ16.00
|
|
4000-7999
|
¢æ11.00
|
|
8000 and more
|
Contact us
|
| License management
|
|
| Volume license manager
|
¢æ16.00
|
| SentiSight 1.1 enterprise license
|
|
| SentiSight 1.1 enterprise license
|
Contact us
|
|
|
|
|
