Intelligent/autonomous real-time applications for detection, target tracking and classifying

Important results from the project

In general terms, the project has generated a deep scientific understanding about visual target tracking. Fundamental insights were achieved about object detection, motion of partly occluded objects, and motion in front of dynamic textures (water). One method was identified as a candidate for an existing product (extended life). The project has implied scientific problem definition, literature surveys, seminars, scientific rethinking where seeds were allowed to grow together with a synthesis with previous knowledge, one conference paper (ICPR16) and one patent application.

Expected long term effects

The most concrete results are formed by a scientific literature survey within the field, one method that could be implemented in a product so that its precision is improved and its life time is extended. The planned patent is applicable in everyday situations so that its segment of application is larger than just target tracking. The forthcoming projects will be able to utilize the knowledge from the project, consisting of specific literature surveys, implemented subroutines, and insights about important principles of the image processing.

Approach and implementation

The relatively limited project was composed by a few fundamental phases: literature surveys, synthesis with our previous knowledge, production of results, and to some extent presentation of the results at an international conference. The project will also consist of a post-phase: one patent application with minor examination, modification and testing. The forthcoming projects could also be viewed as a continuation of this project, in which important approaches and procedures will be applied in a practical context.