-
Notifications
You must be signed in to change notification settings - Fork 4
Journal image tagger. Tcl/Tk and matlab based app for segmenting and tagging document image regions then running various classifiers over them
License
scttl/jtag
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Folders and files
Name | Name | Last commit message | Last commit date | |
---|---|---|---|---|
Repository files navigation
JTAG ==== Journal Image Tagger software developed by Scott Leishman and Kevin Laven circa 2003. A brief overview and a couple of screenshots can be found at: https://www.cs.toronto.edu/~scottl/data/jtag.html 1. Directory Contents --------------------- This directory is the base (root) directory and contains everything related to the Journal image TAGging application software (JTAG). This software is a GUI driven Tcl/Tk application that can load, display, and magnify raster image representations of documents in a variety of formats. Once displayed in the application, the user can add, display, manipulate, and save overlaid rectangular region selections (each of which is intended to isolate one particular document component ex. a page number or section title). In addition to this application, a series of matlab functions and scripts are available to automatically recognize and classify regions on new pages. This is accomplished through an appropriate learner and a corpus of accumulated training data. Directly contained in this directory is the Tcl/Tk source to the JTAG application and a series of subdirectories that aid the application in one way or another (see Section 1.1 for a description of each). Note that there are some dependencies on the JTAG source and certain aspects of the subdirectory structure so it is important that this structure remains intact during its use. 1.1 Subdirectory Contents ------------------------- CVS - (internal / not used) CVS version information bin - linux specific binary executable utilities used by the JTAG application doc - houses all documentation related to the JTAG application lib - 3rd party Tcl/Tk packages used by the JTAG application matlab - holds all matlab functions and scripts used to generate training data and classify page segments scripts - contains various helper scripts to aid in everything from JTAG application installation to document downloading and conversion src - contains tarballs of dependency software (Tcl/Tk, BLT etc.) For further information on each subdirectory, see its own README file. 2. File Contents ---------------- .jconfig - Sample JTAG configuration file listing options that can be set by the user. This file is read by the JTAG application upon startup README - this file classify.tcl - JTAG source file responsible for all things related to rectangular selection regions, and the various classes (buckets) config.tcl - JTAG source file containing all the data shared between the other source files, and the logic to read and write the configuration file file.tcl - JTAG source file responsible for handling all "lower-level" file parsing and writing image.tcl - JTAG source file used to manipulate and display the raster image main.tcl - JTAG executable and the point of source code entry menus.tcl - JTAG source file responsible for all the menu entries and commands ui.tcl - JTAG source file responsible for packaging up the various components for display in an appropriate fashion 3. Installing the JTAG Application ---------------------------------- Since the JTAG application is written in Tcl/Tk (an interpreted language) no compilation is necessary. All you must do is ensure that you have the following available on your local machine: * Tcl/Tk 8.4 (tested with 8.4.3) * TkImg 1.3 (aka Img extensions) * BLT 2.4 (tested with 2.4z) * libtiff, libjpeg etc. All these are included with TkImg To automatically test for and install any missing dependencies from the list above, simply issue: scripts/install_jtag.sh from this directory, and follow the on screen directions given. Note that the only tested platform currently is Linux, but other *NIX platforms should probably be little (no??) effort to port to. 4. Starting the JTAG Application -------------------------------- Should be as simple as running main.tcl contained in this directory. You can also specify the path and name of a valid image file to load at startup as an argument. 5. Using the JTAG Application ----------------------------- Most things should be fairly intuitive. For a complete reference see the file: doc/useage.txt for more information.
About
Journal image tagger. Tcl/Tk and matlab based app for segmenting and tagging document image regions then running various classifiers over them
Resources
License
Stars
Watchers
Forks
Releases
No releases published
Packages 0
No packages published