tina [about] [news] [software] [demos] [projects] [docs] [images] [faq] [people] [teaching] [tina4] [tina5] [WIKI] [KNOPPIX

jump to




Tina 4 is an older, deprecated version of the Tina software. The move to Tina 5, initiated by the Osmia project, has resulted in numerous improvements, including various bug fixes, a new structure for the header files, the split into tina-libs and tina-tools, an autoconf-based build system, a Gtk interface, a CVS repository for the code, and these web pages. Users should migrate to Tina 5 as soon as possible. However, for various reasons (equivalence-testing of algorithms, historical interest etc.) this page provides access to the older versions of the software. Note that we will no longer support Tina 4, and new users should use Tina 5 instead.

Tina 4 has been compiled on a variety of hardware under a range of Unix modelled OS's including, Solaris, Linux and IRIX. It requires a `c' compiler such as `gcc' an X window environment and either Xview or Motif (Lesstif) to work.

Tina 4 is distributed in 2 parts. The first is the library source, containing all you need to build the base Tina 4 libraries. This pack is available below and should be downloaded and installed first. The second part in a toolkit. The library source includes a simple tool example called tinatool which acts as a starting point for your developement. More advanced toolkits (often with example datasets) can be downloaded from the page further below.

tina 4.0.0 [download] [Changes.txt]
tina 4.0.1 [download] [Changes.txt]
tina 4.0.2 [download] [Changes.txt]
tina 4.0.3 [download] [Changes.txt]

Code viewing

The Tina 4 library code can be viewed here using either lxr or doxygen.

tina 4.0.2 [lxr] [doxygen]

toolkits and projects


The demonstrations tools are made available in two forms; source code & binary executable (Sun Sparc Solaris & x86 Linux). In either case each demonstration is packaged as a tarball - tar'd and zipped (*.tar.gz).

In order to compile the source-code it is necessary to have an existing version of the base TINA libraries available from this page. Once you have the libraries installed the demonstration tool may be downloaded, compiled and executed. Once the toolkit has been unzipped and untarred into a convenient location, cd into the directory and type "make". Each toolkit contains a README file with instructions on how to use the software.

If you do not want to install a version of the TINA libraries you can download an executable version of the demonstrations. To execute these you will need either a Sun Sparc Workstation running Solaris 2 (version 2.7 of Solaris was used to compile the Sun code). Or a x86 compatable PC running a version of Linux (Linux kernel 2.2.5 - glibc 2.0 from SuSE 6.1 was used to compile the Linux code). If you do not have either of these systems you will need to compile a version for yourself.

Machine Vision
Several interactive demonstrations of machine vision algorithms.
[The Image Calculator] RPN (stack based) calculator working on images.
[2D Object Recognition] Pairwise Geometric Histogram based approach to robust model based 2D object recognition and location.
[3D Wireframe Model Matcher] System for matching wireframe models to 3D edge data extracted using auto-calibrating stereopsis.
[Automated Stereo Camera Calibration] Robust approach to stereo camera calibration using unconstrained scene data.
[Temporal Stereo Depth Estimation] Feature based stereo reformulated as an area based algorithm with temporal extensions for greatly improved robustness.
Medical Image Analysis
Demonstrations of algorithms with a more medical application (although these algorithms may be useful in a wider context) are presented below. Note that this software and associated data are supplied as illustrations of research analysis procedures documented in the public literature and not as clinically proven techniques. Many of these demonstrations have been combined into the same download package (check the filename before downloading more than one... you might have it already!)
[MR Coil Correction] Automated MR coil non-homogeneity correction.
[Image Coregistration] Manual and automatic volume realignment and fast sync reslicing.
[fMRI Motion Artefact Removal] Measurment of residual motion artifacts in realigned fMRI datasets.
[Deformable Region Extraction] A trainable model approach to automated deformable boundary location.
[MR Tissue Segmentation] Probabilistic partial volume tissue segmentation.
[MR Perfusion Analysis] Blood flow analysis from contrast enhanced MRI.

Valid HTML 4.01!