IEEE Transactions on Pattern Analysis and Machine Intelligence
Deformable Contours: Modeling and Extraction
IEEE Transactions on Pattern Analysis and Machine Intelligence
IEEE Transactions on Pattern Analysis and Machine Intelligence
Statistical models in medical image analysis
Statistical models in medical image analysis
Pattern Recognition and Machine Learning (Information Science and Statistics)
Pattern Recognition and Machine Learning (Information Science and Statistics)
Snakes, shapes, and gradient vector flow
IEEE Transactions on Image Processing
Active contours for tracking distributions
IEEE Transactions on Image Processing
| Hi-index | 0.00 |
Automated tracking of deformable objects that change shape and size drastically is challenging. For useful results, one needs an efficient deformable object model. In this regard, we propose a novel deformable object model via joint probability density of level set function and image intensity/feature values. Given the delineated object boundary on the first image frame of a video sequence, we learn the aforementioned joint probability density via kernel (Parzen window) method. From the next frame onward, we match this learned probability density with the probability density on the current frame by minimizing Kullback-Leibler divergence. This minimization procedure is cast in a variational framework and a minimizer is obtained by solving a partial differential equation (PDE). A stable and efficient numerical scheme is proposed for solving this resulting PDE. We demonstrate the efficacy of the proposed tracking method on myocardial border tracking from mouse heart cine magnetic resonance imagery (MRI).