Forward Vehicle Tracking Based on Weighted Multiple Instance Learning Equipped with Particle Filter
Keunho Park, Joonwhoan Lee · Journal of Korean institute of intelligent systems · 2015
Abstract Ths piape prorposes a nov elofwradr vehec tliarcknig agolhtmir based on hte WMLIW(geihetd Mpeiul tlInstance Learning) equipped with a particle filter. In the proposed algorithm Haar-like features are used to train a vehicle object detector to be tracked and the location of the object are obtained from the recognitionresult. In order to combine both the WMIL to construct the vehicle detector and the particle filter, the pro-posed algorithm updates the object location by executing the propagation, observation, estimation, and se-lection processes involved in particle filter instead of finding the credence map in the search area for every frame. The proposed algorithm inevitably increases the computation time because of the particle filter, but the tracking accuracy was highly improved compared to Ababoost, MIL(Multiple Instance Learning) and MIL-based ones so that the position error was 4.5 pixels in average for the videos of national high-way, ex-press high-way, tunnel and urban paved road scene.Key Words : WMIL, Particle Filter, Forward Vehicle, Object Tracking, Video Frames, Haar-like FeaturesReceived: Jan. 26, 2015Revised : Jun. 19, 2015Accepted: Jun. 22, 2015