A Robust Parametric Active Contour Model for Target Tracking using Modified Energies: Virtual Electric Field and Motion-based Balloon
Publish place: 19th Iranian Conference on Electric Engineering
Publish Year: 1390
نوع سند: مقاله کنفرانسی
زبان: English
View: 1,303
This Paper With 6 Page And PDF Format Ready To Download
- Certificate
- من نویسنده این مقاله هستم
استخراج به نرم افزارهای پژوهشی:
شناسه ملی سند علمی:
ICEE19_099
تاریخ نمایه سازی: 14 مرداد 1391
Abstract:
Active contour model (ACM) is a powerful tool for the target tracking in digital image sequences. Traditional ACM fails to track the targets when target displacements or aspect changes are high and inhomogeneous. In order to improve these inefficiencies, in this paper, a modified balloon energy in the parametric ACM is defined base on target displacement in two successive frames to inflate some suitable points adaptively and locally, in the current frame. Moreover, traditional ACM suffers from low capture range, so it can not attract to the concave boundaries of the target those results to an uncertain tracking in image sequences. To improve this inefficiency, a modified virtual electric field energy is used in conjunction with the proposed balloon energy. Also, new ACM algorithm adapted base on mutual changes of target and background gray levels. The advantages of the proposed ACM consist of: less sensitivity to initialization, large capture range, attraction to sharp-pointed and concave boundaries of target, ability of tracking the target with large and inhomogeneous displacements, and acceptable computational cost. Experimental results show that the tracking by the proposed ACM based on the greedy algorithm produces more correct detection percent of the target boundaries than other similar methods, in different circumstances
Keywords:
Target tracking , active contour model(ACM) , motion-based balloon energy , virtual electric field energy
Authors
Mahdi Peymani
Department of Electrical Engineering, University of Isfahan, Isfahan, Iran
مراجع و منابع این Paper:
لیست زیر مراجع و منابع استفاده شده در این Paper را نمایش می دهد. این مراجع به صورت کاملا ماشینی و بر اساس هوش مصنوعی استخراج شده اند و لذا ممکن است دارای اشکالاتی باشند که به مرور زمان دقت استخراج این محتوا افزایش می یابد. مراجعی که مقالات مربوط به آنها در سیویلیکا نمایه شده و پیدا شده اند، به خود Paper لینک شده اند :