Robust Multiresolution Estimation of Parametric Motion Models Applied to Complex Scenes Jean-marc Odobez and Patrick Bouthemy
نویسندگان
چکیده
This report describes a parametric motion model estimation algorithm. Motivations for the use of such models are, on one hand, their representativeness, which has proved to be eecient in numerous issues such as measurement, segmentation, tracking and interpretation of motion, on the other hand, their compactness and, therefore, the low computational cost of the estimation of these models. However, it is important to have the best accuracy for the estimated parameters, and to take into account the problem of multiple motion. We have developed two M-estimator-like techniques in a multiresolution framework. Numerical results support this approach, as demonstrated by the use of these algorithms on complex sequences. Estimation robuste multi-echelle de mod eles param etr es de mouvement sur des sc enes complexes R esum e : Nous pr esentons dans ce rapport une m ethode originale d'esti-mation de mod eles param etr es de mouvement. L'int er^ et de formuler un pro-bl eme d'analyse de mouvement par l'identiication de tels mod eles est double. Il s'agit d'une part d'une repr esentation compacte qui s'av ere ad equate et pertinente dans des contextes dii erents et nombreux (mesure, segmentation, suivi ou caract erisation du mouvement). D'autre part, on peut en obtenir une estimation peu co^ uteuse en temps calcul. Le point crucial par contre est d'en obtenir une estimation able et pr ecise, et de g erer correctement la pr esence de plusieurs mouvements dans l'image. Pour r epondre de faa con bien fond ee et eecace a ces probl emes, nous avons d evelopp e un estimateur robuste, du type M-estimateur, d eeni dans un sch ema multi-r esolution. Cette technique a permis d'obtenir des r esultats tout-a-fait satisfaisants sur des images re-pr esentant des sc enes complexes.
منابع مشابه
Robust Multiresolution Estimation of Parametric Motion Models
This paper describes a method to estimate parametric motion models. Motivations for the use of such models are on one hand their efficiency, which has been demonstrated in numerous contexts such as estimation, segmentation, tracking and interpretation of motion, and on the other hand, their low computational cost compared to optical flow estimation. However, it is important to have the best acc...
متن کاملMRF-based motion segmentation exploiting a 2D motion model robust estimation
This paper is dealing with motion-segmentation, that is, with the partitioning of the image into regions of homogeneous motion. Here, homogeneous means that in each region a 2D polynomial model (e.g. an aane one) is able to describe at each location the underlying \true" motion with a predeened precision. However, no estimation of this true motion eld is required. The motion models are computed...
متن کاملDirect incremental model-based image motion segmentation for video analysis
Dynamic analysis of image sequences is an important task in object-oriented video applications. It often relies on the segmentation of each image of the sequence into region entities of apparent homogeneous motion. In this paper, we present an original motion segmentation algorithm based on 2D polynomial motion models, a multiresolution robust estimator to compute these motion models, and appro...
متن کاملReal-Time Estimation of Dominant Motion in Underwater Video Images for Dynamic Positioning
In this paper we propose a D visual motion estimation method which can be exploited to achieve a dynamic position ing eg by gaze control with respect to a sea bottom area of interest of a video camera mounted on a subsea vehicle It mainly involves a dominant D motion robust estimation step from underwater video sequences Optimizations car ried out on the motion estimation code have made possibl...
متن کاملRobust multiresolution computation of optical flow
This paper presents an algorithm that provides good estimates of the optical flow at motion discontinuities and occlusion regions based on the least-median-squares robust estimator. The use of a multiresolution scheme enables optical flow to be computed for objects with large motions. A movable overlapping neighbourhood strategy eliminates block-effects and also deals with cases of the local ne...
متن کامل