software

The server enables upload of images and video sequences, which will be analysed consecutively. Videos will be decomposed into sequence of images first, which will processed. The image processing is divided into several modular phases (segmentation, segment classification, object detection), which can be added or modified as needed. It is possible to query the analysed data, e.g. in following ways:    - Does the image/video contain given type of object, e.g. car, pedestrian,      traffic sign etc?    - Does the image/video contain given type of area, e.g. grass, asphalt, sky,      fire etc?    - Are predefined relations between objects and areas in the image/video valid?      E.g. adjacency, presence etc.      The system also allows to define so-called "tags" in the form of predefined      queries and logical combinations of them. An image/video is tagged if it      satisfies defined query.      Services provided by the server are useful in e.g. crisis management, when      it is useful to sort images and videos according to predefined (crisis)      scenarios.      Implemented in Python, C, C++ and XML. The target operation system is Unix      (must include Python 2.6 interpreter, Twisted network framework and C/C++      compilers). The server-client communication uses HTTP protocol, the server responses      are either images/videos or XML stuctures.

Image Processing, Scene Interpretation, Image Segmentation, Segment Classification, Object Detection

15. 12. 2010

http://www.fit.vutbr.cz/research/view_product.php.cs?id=165

K využití výsledku jiným subjektem je vždy nutné nabytí licence

Poskytovatel licence na výsledek nepožaduje licenční poplatek

http://www.fit.vutbr.cz/research/view_product.php.cs?id=165