A Real-Time Evolutionary Object Recognition System

We have created a real-time evolutionary object recognition system. Genetic Programming is used to automatically search the space of possible computer vision programs guided through user interaction. The user selects the object to be extracted with the mouse pointer and follows it over multiple frames of a video sequence. Several different alternative algorithms are evaluated in the background for each input image. Real-time performance is achieved through the use of the GPU for image processing operations. 1 Motivation Current vision systems are usually not adaptive to their environment. Algorithms which have been developed in the laboratory often break when the vision system is moved to a different environment. The only component of a vision system which is currently adaptive is the automatic white balance of the camera. Recently, Ebner [1] proposed building an adaptive on-line evolutionary visual system. Such a system would adapt itself to the current environment. Evolutionary algor...

• Type: Journal
• Year: 2009
• Where: EUROGP
• Authors: Marc Ebner
• Keywords: Algorithm, EUROGP 2009, Image, Main Algorithm, Optimization