Copyright © 1999-2011 LTU Technologies.
|Revision $Revision: 1.33 $||$Date: 2011-03-24 17:39:32 $|
The document has been designed for users of LTU engine 5.0. The goal of this document is to explain in detail the user functionalities of the browser-based interface.
The document has been designed for users of LTU engine 5.0. The goal of this document is to explain in detail the user functionalities of the browser-based interface.
You can enter the LTU engine web-based interface on your server at the following URL:
The following table summarizes the minimum client system requirements to access LTU engine 5.0.
Table 2.1. Client system requirements
|Operating System||CentOS 5, Debian 5, Debian 6, Windows server 2003 and Windows server 2008. Support by default 64bits Operating systems.|
|Browser||Microsoft Internet Explorer 6.0 or higher, Firefox 1.5 or higher.|
|Java Plug-In||Sun Java 2 JRE, Standard Edition v 1.4.2 or newer (1.5 recommanded).|
For security reasons access to this application is protected. Each user has to log in at the beginning of a session.
During a session if you need to log as another user, click on:
This will initiate another session.
After 30mn of inactivity, you will have to log again (same session). This setting can be changed.
When a user logs in the system, the homepage presents default access to the last accessed repository. Additional information on this repository is provided, for instance on the user's projects relative to this particular repository.
To avoid user confusion on the repository that is currently accessed, the home page is the only page where the user can change repository.
The menu may not offer all the choices based on the user rights, defined in
The top of each page includes a menu that allows navigation between the main features.
Reports > Autoscan is a dynamic menu which is created when an Autoscan search report is available.
My Workspace (not available by default) > My Projects is a dynamic menu which is available when a project has been created.
The homepage provides a synthetic view of main options. Click:
to get the view.
Compared to previous releases, LTU engine now comes with many new icons that are easy to remember. Here are tables that summarize all correspondences.
To display the icons on the image bar, you must drag the mouse on it.
Table 2.3. Icons
|Display image properties|
|Bouncing with color|
|Bouncing with color and shape|
|Bouncing with shape|
|MD5 presence indicator|
|IPTC metadata presence indicator|
|EXIF metadata presence indicator|
Table 2.4. Icons (cont.)
|Quit LTU engine||Edit|
|Select in page||View as...|
|Copy to clipboard||Refresh|
|Save in project...|
|Quick search||Export as...|
|Go back||Go back to action of the list|
|Go forward||Go forward to the action of the list|
The user interface integrates navigation facilities for going back to previous significant user actions of the workflow and going forward to those actions that follow an action coming from the history list.
Once the user has made an action from the menu of the UI, the back button becomes active and permits going back to the previous user action (one step back or several steps back in case the user did more than one action). All previous actions compared to the current page are put into a list in reverse order (the latest action on the top). When the user clicks on one of the actions of this list, the list is actualized according to the current page that changes.
The below diagram illustrates entries of user actions that are in the history list of the back button:
Once the user has used the back button or has taken an action from the associated list to go back to an action, the forward button becomes active. Depending on the position of the step the user went back regarding the whole action sequence, it is possible to go one or several steps forward with the forward button. All actions that follow a given action compared to the different navigation steps the user did are put into the list of forward actions.
Here is an example of user actions in the list of the forward button:
Everytime the user executes an action from the menu bar, the actions of the forward action list are put into the list of previous actions related to the back button. This behaviour is different from the behaviour of an internet browser where the list of forward actions is reinitialized.
Information about system and repository configuration is accessible from the menu
System > Configuration.
Information about system configuration concerns the database, the license and global settings.
Information about repository (image database) configuration concerns parameters related to the signature and the thresholds that have been defined.
The following pieces of information are given:
identifier of the repository
name and type of image signature (DNA)
activation of segmentation or not
thresholds for retrieval, clone, similarity and video matching
storage directories for signatures and images
LTU engine enables to define multiple users, profiles and groups with diverse rights to diverse image databases(repository).
User, profile & group terminology
A User is an individual, that is allowed to use LTU engine.
A Profile is the set of rights assigned to a user. On each type of object (each view) in any repository, the Profile assigns read/write rights. These rights apply to group objects or all objects.
A Group is a community of users (to be defined in
An object is created by a user. It belongs to the group of the user, but this can be changed. User rights depend on profiles (not on groups) : depending on its profile, a user can see all objects or only the objects that belong to its group. If he can see all objects, he might :
edit all objects
or edit objects of its group
or not be able to edit anything.
Some users may have all rights (administrator profile, for example, which is a "Super User").
Group management : it is forbidden to delete a group. The explanation is that group data is not stored in the main database, but in separate files. This point will evolve in future releases, but for now, please define your groups carefully.
LTU engine's databases are separated in terms of image DNA metadata structure and retrieval. Different users in relation with individual user rights can access them. For instance, user A can have access to repository 1, 2 and 4 whereas user B will only access repository 1 and 3.
For each view, administrator can define precisely the edit / browse rights of users on objects.
Image enrollment can be done with image enrollment tool from LTU engine User Interface if the number of images to be enrolled at one time does not exceed the number of 10.000. Otherwise, you can use a batch tool made available by LTU (
[install_dir]/tools/perl). For further information, please contact ltu support at
This page allows the user to enroll images from a directory and its sub-directories. It is organized in three sections:
Selection of the image directory
Form for information fields
Status of enrollment process
The Image Enrollment function is available as a Java application. Click on
Launch Enrollment Application. If you don't have pre-installed the application, you can download and install it from :
Access page to the Image Enrollment Application:
This application uses Java Web Start 1.2. This package must be installed on your machine.
For security reasons, LTU engine will ask you for your login/password. Since adding images is critical for the integrity of the LTU engine content, such login will enable to check which database you can access as a content administrator.
The first step of the enrollment process is to select a directory of images for enrollment. This is done by selecting the "..." button at the top of the page and selecting the desired directory.
1 You can also directly type the directory path in the edit box.
2 The default installation of the LTU engine server contains a maximum image file size limitation. The default value is 17 megabytes, any images over that will not be indexed unless the setting is updated (please notify your database administrator to increase 'http.max.post.size' in ltu_global_conf table).
After the directory is selected, hit
OK and then the
Explore button to load the images in the java applet.
The application can load all images in subdirectories of the selected directory by checking the recursive check box.
During the exploration process, the applet will search for images already known by LTU engine. If an image is already in the image database, the image attributes and pre-defined text fields about this image are pre-entered in the corresponding fields for the new enrollment.
Select the image directory for image enrollment:
Once the images are loaded you can enter visual and textual information for these images.
A thumbnail of the current image is displayed at the top left of the form (the supported image format are: jpg, jpeg, bmp, png, gif, tiff, tif even though not all of them are displayable). The name, the rank of the current image and the total number of images are also displayed with the thumbnail. To navigate through images, you can use the arrows or the "jump to" box. Hit the right arrow to go to the next image, the left arrow to go to the previous image, the double right arrow to go 10 images forward, or the double left arrow to go 10 images backward. Enter a number in the
jump to box and hit
Enter key to directly go to this image. To deselect an image from enrollment, select the
Do not enroll this image check box.
For each image you can fill fields
like Author or Category.
Certain fields are unique to each image; they appear in italic. This means that there cannot be multiple occurrences of the same image with the same value for these fields. If you do enroll an image with such a combination, the information will simply be replaced (you can see this information in the status bar at the bottom of the page during the enrollment).
It is possible to apply a field value to all images by checking off the check box in front of each field (visual and textual). Some fields can be automatically pre-entered.
Image attributes and metadata information form:
When you are ready to enroll all the images, select
File > Enroll.
The enrollment process responds differently depending on the images encountered:
When the image is new to the LTU engine, it is added along with its visual and textual information.
If the image already exists in LTU engine and the unique fields entered have the same values as the existing one, the textual information are replaced.
And finally, if the image already exists in LTU engine but the unique fields entered have different values from the existing one, its image attributes are updated and its textual information is completed with the new one.
A progress bar displays the amount of images treated. Under the progress bar a status line gives more information:
Number of new images enrolled.
Number of new record added (same image, different mandatory fields).
Number of update record (same image, same unique fields).
Number of errors. Due to network problem for example.
The application will flag the user and will not enroll any image if the mandatory information fields are empty.
The user is not allowed to modify data during the enrollment. The only allowed action is the
Stop button that stops the image enrollment (Already enrolled images remain in LTU engine).
If you want to delete an image, you have to prior delete its associated metadata. It is, in fact, the reverse process of enrolling images.
This action is available from any result or navigation page where the symbol is displayed near an object.
The following feature is only accessible to users with content administration privileges.
A new window will open that enables to view the complete information related to the object and to modify it.
The following icon enables to open a printer-friendly browser window:
Two terms are introduced here:
Object type: type of data handled by the relational model of the LTU engine database (e.g. Author).
Object: individual content of an object type (e.g. Name of an Author).
File enables 3 main actions on object types:
Create new objects
View existing objects
View unused objects
An object with an image is the central type of object within LTU engine. In the presented data model, it is called an Asset. The Asset features an image with associated textual information. Such textual information can either be free text related to the image or another object coming from the data model.
Click on the following button to edit object:
The views can be defined in any manner upon configuration of LTU engine.
There are several views possible (list or in details). The access panel to these views :
The user has the possibility to manipulate several objects at once. It is possible to edit several objects in a pop-up window where the user select the attribute to set in all selected object. The same mechanism is used for massive deletion. These features are eased by the
Select all in set checkbox that complete the
Select all in page one.
Example : select images in page. Click on
find related record icon. Then click on to
edit object icon. To apply data changes to all selected objects:
A confirmation message before proceeding (deletion or editing);
A progress bar will show you the process;
If any error, you will get explanations :
For each of object types, the menu allows to create objects from the interface.
The creation of object is automatically reflected in the image enrollment application, thus allowing using the richness of the data management module in the enrollment process.
Creating objects brings to LTU engine the knowledge of the data structure and the data relationship.
Upon installation, it is recommended to design the data model so that the creation of objects is intuitive for content administrators (e.g. relative position of each object type on the menu).
An image has got associated metadata, in normal ways. Associated metadata can be erased or simply not existing (no investigation data) so that the image has no more associated data. It can be interesting in the investigation process to put in evidence the images that have no associated metadata (for example : no investigation yet).
An object needs to be unused to be deleted. Otherwise, the relationships it has with other objects prevent the deletion.
Click on the icon of magnifying glass in the top navigation bar opens the file browser. This permits you browsing folders and selecting an image that you can upload for an image search. Click on
Open/Ouvrir uploads the selected image to LTU engine server and processes an image search.
The image search section provides :
pure visual search capabilities without using the richness of the data management module.
metadata search, combined or not with image search.
This combined text and visual search allows expert searches within the content.
You can easily name a query, and add it to a project.
The saved query can be accessed from the homepage:
The menu enables to browse and navigate through the LTU engine database. It gives users an idea of the content that has been enrolled. From there, it is possible to launch image search queries within the database and sort the content by similarity. For example, select
File > Browse > Photo. Find an image and find similar:
The search result page presents the list of matching objects (textual) and ranks the image database by similarity with respect to the query (visual).
LTU engine memorizes the last user actions. When switching from one page to another, the query summary is always accessible to the user.
Add to clipboard option:
Click once on the image to select it .
The principle of My Workspace is to provide users with the capability to capitalize on their searches and use of LTU engine to store, comment and share information about retrieved content.
By default, My Workspace is not available, please contact LTU to activate it.
A project is a user workspace to store images and searches. Many projects can be created, commented and shared through printer-friendly interface. It is possible to switch from one project to another.
Here is an example of a project that contains image results. There are no queries in this report.
My Seeker remembers the queries that are performed by a user. They are stored in My Seeker with summary information and possibility to re-run them. Queries can be found also in
Objects > Search history.
User configurable reports: You can add comments on any item (query or object) of a project, reorder the objects and print the projects with the ordered queries and objects on the same page. You can also delete objects in the project only or in the database.
User queries are saved in the search history and accessible from My Workspace > Search History:
The different queries can be re-run by clicking on the corresponding icon.
You can select objects (images, records, cases, etc.) and put them into the clipboard with the following icon on the main navigation bar:
The objects of the clipboard are accessible from
My Workspace > Clipboard. A tab is created for each object type. In the example below, images, records and cases have been put into the clipboard. The active tab concerns images.
With LTU engine, LTU AutoScan lets the user select many files and folders that will be submitted against the enrolled images of an LTU engine system. A detailed report containing visual similarities information is then generated. User can run multiple AutoScan queries.
Three levels: duplicate, clone and similar.
Minimum client system requirements: Windows 2000, XP or Linux, IE 6 or N7.
Autoscan detects hidden files, with false extensions (like .ppt .exe .doc .dll , and many more).
Download to install AutoScan application. Save the file and execute.
Table 7.2. Autoscan installation
|1 Start installation. Click "next"||2 Install|
|3 Installation progress...||4 Finish|
In your windows
Program Files, select :
LTU engine AutoScan > or from LTU engine User Interface
Tools > AutoScan Application.
Table 7.3. Autoscan application
|Program starts.||For security reasons, authentification is needed.||Choose workspace.|
Here are the steps for creating an AutoScan search report from an uploaded image folder.
Browse your drives to find your images.
Select your images in the left column (
Ctrl for multiple selections).
Add your images in the right column, using either or
Actions > Explore Folder. Your images are then in your basket in the right column :
In replacement of step 2, you can also explore folders recursively. Select a folder:
Actions > Explore folder to find images. A dialog box displays asking you if you want to search for images in sub-folders.
You can access the thumbnail preview and attached metadatas.
Select images you want to submit against LTU engine (Individual selection or
Actions > Select all images).
Upload selected images from
Actions > Upload images or click on the corresponding icon.
Enter a name of the report and a description (optional). You can also select filters that will be used for defining the search perimeter in the image database. For example, if you select the filter
With victims, the uploaded query images will be compared only to images in the database that have meta-data related to
AutoScan then uploads all images on the server :
In the following are described the steps in order to create an AutoScan report from a XML file. The XML file is a file that has been exported from an LTU engine database as described in chapter 10.2.
Select in the menu
Actions > Create a report from a XML file:
Browse the file system and select a XML export file. In the right pane of the windows, file properties and content are given. After having selected and opened the XML file, the search task for detecting duplicate, cloned and similar images is launched. The result will be available in the AutoScan report.
The progress bar informs you about the upload status.
After upload of the images, an AutoScan search task is performed on server side. The completion of the search task can be viewed from LTU engine user Interface.
Go to LTU engine, in
Reports > Search Reports Status you can display the status of the search report.
The status of all AutoScan search reports is displayed as follows:
All available search reports are included in the menu
Reports > AutoScan.
You can configure in
File > Preferences the number of thumbnails is limited by the power of the hardware configuration.
You can search for image matches in LTU engine database in comparison to a set of MD5 hashes that you can upload to LTU engine. You can access to this functionality from
Tools > MD5 Search. This menu is not available by default, please contact LTU to activate it .
The MD5 search form presents as follows:
The format of the MD5 file to be uploaded is 1 MD5 per line. This type of file can be provided by an export of MD5s coming from an LTU engine database. The export can be done by a volume export of MD5s as described in chapter 10.1.
b4d77325e5bd4479f706eeb821bc373b b27f34afde354b5a298f0c33ed6f76d4 d7a0604fae116862c0d8c47de37afae6 e3265d55094f9a808f8eaa0ffe193ebb
After having uploaded the file with the MD5 hashes, a task is scheduled on LTU engine server side. You can follow the percentage of completion of the task after refresh of the
MD5 Search page.
As a result, you can access to two tabs, one for matches and the other for unmatched images. Matches are images that exist in LTU engine with a MD5 that figures in the external file that has been uploaded. Unmatched images correspond to MD5 hashes in the external file that do not have an equivalent representative in LTU engine database.
You can get more detailed information for the results by click on the
view object detail icon:
You can export a project in different formats. The following ones are available: LTU format (specific XML schema), zip archive, Microsoft Word and Microsoft Excel files.
Download is ready when the file in the selected format has been generated.
Below is an illustration of an export of a project as Microsoft Word document.
You can export the integrated AutoScan report as files in LTU format or as a zip archive, Microsoft Word document or Microsoft Excel file.
After you have chosen an export format, the export is processed and can be downloaded as a file after the file generation has finished.
The html page of an AutoScan report included in the zip archive looks as follows:
Here is an example of an AutoScan report exported as Microsoft Word document:
The result of a MD5 search can be exported in LTU format, as a zip file, as a Microsoft Word document or Microsoft Excel file.
Below you can find an example of the export of MD5 search in the XML format of LTU.
Here is an example of the html page included in the zip file of a MD5 export:
The Microsoft Word document of a MD5 export presents as follows:
A repository may be configured with volume export functionality. This functionality is available by default . As suggestied by the name, volume export permits to export high volume of data.
Volume export permits to exchange data:
exchange of MD5 data,
exchange of image data and related meta-data in a XML file.
MD5 export results as well as export results in XML format can be used for search tasks in LTU engine:
MD5 export file can be used for MD5 search.
XML export file can be used for Autoscan search task.
XML export file can also be used by the software LTU engine for EnCase, the alliance of Guidance Software's EnCase and LTU engine of LTU, for photographic evidence analysis on seized or intercepted materials. For further information, please contact
MD5 export exports the MD5s in a text file containing one MD5 per line.
This functionality is accessible from
Tools > Volume Export > MD5 Export Form:
Using the export form, the user filters out the set of MD5s to export. In the example below, he may decide to export MD5s from the images of a particular case et enrolled a particular date.
XML volume export is accessible from
Tools > Volume Export > XML Export Form:
The principle of XML export consists in applying an export profile to whole or part of the database content for exporting it. The export profile defines which object(s) will be exported.
The XMl export form has three sections:
Export parameters: Naming the exported file and associate a description.
Export profile: Access to export profiles for creation / edition / deletion or selection.
Select content to export: Search fields for selecting the content to export.
It presents as follows:
Sig is an example of profile that has been created by a user.
When you clic on
Create a new profile, you access to the following form. Some attributes have been selected as an example.
After you have launched the export from the
XML Export form, an export task is scheduled on server side. After refresh of the
Export Task page, you can follow the progress status of the task.
After completion of the task, you can download the file containing the result. Here is an example of a XML report:
System > Diagnostic Tool gives you access to a page on which you can select which log and configuration files you want to download from LTU engine for diagnostic and troubleshooting purposes. Clic on
Diagnostic downloads one archive file with the requested files.
In case of problems you may encounter or on request coming from LTU's support team, please send the downloaded archive file to
Online help is accessible from the menu
? > Help for the following documents:
Content Administration Guide
Task Export Configuration Guide
Frontend Configuration Manual
A html version for online consulting and a printable html version are available.
LTU engine server 5.0 can be used to search on colors. First, create a repository selecting the "Color Query" template.
Then, give the appropriate authorizations to your repository, select it and click on the menu "Samples/Color Query Demo".
You have now access to a precomputed palette. Click on a color to search it.
Anytime you can check "Dynamically Computed Palette" to compute dynamically a palette based on the colors of actual matching images. If you check it before any search, it will display the dynamic palette of all repository images.
After a first click on a color, you can affinate your search by clicking on a new color. You can reset anytime to start a search on new colors. Below you can see the result of an orange and red search. You can search on the image colors represented under the image by directly clicking on it.
LTU engine server 5.0 can be used to perform fine images comparisons. Click on the menu "Samples/Fine Comparison".
It can be used to compare images which are not yet stored in repositories. The script /home/ltu/tools/perl/compareImage.pl is available to compare images already stored in the repository (launch it with -h option to display usage).
We have here compared two almost similar pant images to show their differences. On the left, you can see the resulting error map, on the right, the overlapped images and below the spotted errors displayed in a new window.