Developed by Daniel Rodríguez Molina
Advised by Manuel J. Marín Jiménez
Overview
|
The software released here implements a C++ library for object detection in images. The main algorithm is based on the detection method proposed by Felzenszwalb et al. [1]. The input of the detector is a single image, where the detection procedure is carried out. Depending on the target object class, a different object model is used. Such model is a computational description of an object class. It is stored in a binary file with .mat extension. Some of these model files are included in the released package. |
|
|
LibPaBOD is built on top of two libraries: OpenCV [2] and MatIO [3]. Among other things, OpenCV library is used to handle images using the IplImage structure. In addition, CvMat and CvMatND structures are much more functional in order to perform matrix operations. MatIO library is a basic point. It allows that LibPaBOD can read the object model file and load it into memory. |
||
Example of results
The next figure shows some examples returned by the library after performing the detection process. Click on an image to enlarge it.
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
Downloads
| Filename | Description | Size |
|---|---|---|
| LibPaBOD at GitHub | Source code available at GitHub | -- |
| LibPaBOD for MS Windows at SourceForge | Executable, dll and lib for object detection that runs on MS Windows, built on top of LibPaBOD | -- |
| upperbodyfrontal4libpabod.zip | Frontal upper-body [4] model ready to be used with libpabod. | 95 KB |
|
libpabod_v0.2.4_lite.zip [OBSOLETE: best download from GitHub] |
Contains the full library: source code, example images and models. | 3.25 MB |
| README.txt | Description of contents. | 3 KB |
| libmatio150_bin_win.zip | Lib and dll of libMatio to be used on MS Windows with libPaBOD | 216 KB |
Related Publications
- [1] P. Felzenszwalb, R. Girshick, D. McAllester, D. Ramanan. "Object Detection with Discriminatively Trained Part Based Models." IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 32, No. 9, September 2010. URL
- [2] OpenCV library reference. URL
- [3] MatIO library reference. URL
- [4] VGG's upper-body detector. URL