Journal of Digital Information Management ZHOU Jia-Nan, FENG Zhi-Lin, LIN Zi-Huai 13 6 2015 http://dline.info/fpaper/jdim/v13i6/v13i6_1.pdf Accurate and robust tracking of defect targets in dynamic ink-jet printing videos is a huge challenge in ink-jet printing technology for digital fabrication, and becomes a popular topic for digital information applications in ink-jet printing industry. Recently, particle filters have drawn a significant amount of interest because of their robust tracking performance. In this paper, a robust defect tracking algorithm for ink-jet printing fabric products was proposed in a particle filter framework. First, the ink-jet printing defect tracking was regarded as a Bayesian estimation problem, and a representative dynamic state model was then defined prior to processing the video data. Second, an observation model based on color histogram was introduced to calculate the likelihood of sample particles. Finally, a new motion constrained resampling rule was designed by which the proposed algorithm can track defect targets undergoing abrupt motion conditions and background changes. Experimental results demonstrate the effectiveness and superiority of the proposed algorithm on tracking defect targets undergoing various challenging conditions.