一種新的基于時(shí)空馬爾可夫隨機(jī)場的運(yùn)動(dòng)目標(biāo)分割技術(shù)
A Novel Moving Object Segmentation Technology Based on Spatiotemporal Markov Random Field
-
摘要: 在圖像處理領(lǐng)域，視頻圖像序列中的運(yùn)動(dòng)目標(biāo)分割技術(shù)是一個(gè)被廣泛研究的熱點(diǎn)課題。該文提出一種新的基于時(shí)空馬爾可夫隨機(jī)場的運(yùn)動(dòng)目標(biāo)分割技術(shù)。首先，對(duì)視頻序列的前后3幀圖像進(jìn)行處理，獲得兩幀初始標(biāo)記場；隨后，對(duì)兩幀初始標(biāo)記場進(jìn)行與操作，獲得共同標(biāo)記場；最后，以原始圖像的色彩聚類圖像作為先驗(yàn)知識(shí)，重新定義Gibbs能量函數(shù)，并利用迭代條件模型（ICM）實(shí)現(xiàn)最大后驗(yàn)概率（MAP）的估算問題，獲得優(yōu)化標(biāo)記場。實(shí)驗(yàn)結(jié)果表明：該模型克服了傳統(tǒng)時(shí)空馬爾可夫隨機(jī)場模型因運(yùn)動(dòng)產(chǎn)生的顯露遮擋現(xiàn)象，同時(shí)減弱了運(yùn)動(dòng)一致性造成的空洞現(xiàn)象并削弱了噪聲的影響。
-
關(guān)鍵詞:
- 圖像分割;馬爾可夫隨機(jī)場;迭代條件模型
Abstract: In the field of image processing, the segmentation of moving object in video sequences is a hot research topic in recent years. In this paper, a novel method of moving object segmentation based on spatiotemporal Markov Random Field(MRE) is proposed. Firstly, two observations and two initial labels are derived from the three successive images with the same method in the first scheme. Secondly, the AND-label is obtained with the AND-operation on the two initial labels. Finally, the image segmented with the color clustering algorithm is regarded as prior knowledge, with which the corresponding Gibbs energy function is redefined, and the maximum a posteriori estimator, which is determined by using the iterated conditional mode algorithm, is employed to get optimized labels. The new MRF model contributes to the weakening of the noise and to the elimination of the covered-uncovered background and to the recovery of the uniform moving regions. -
Kim Munchurl, Choi Jae Gark, Kim Daehee, et al.. A VOP generation tool: Automatic segmentation of moving objects in image sequences based on spatiotemporal information[J].IEEE Trans. on Circuits and Systems for Video Technology.1999, 9(8):1216-[2]Fan Jianping, Yu Jun, Gen Fujita, et al.. Spatiotemporal segmentation for compact video representation[J].Signal Processing: Image Communication.2001, 16(6):553-[3]Meier T, Ngan K N. Automatic segmentation of moving objects for video object plane[J].IEEE Trans. on Circuits and Systems for Video Technology.1998, 8(5):525-[4]詹頸峰，戚飛虎，王海龍. 基于時(shí)空馬爾可夫隨機(jī)場的運(yùn)動(dòng)目標(biāo)分割技術(shù). 通信學(xué)報(bào)，2000，21(11)：120.126.[5]Luthon F, Caplier A, Lievin M. Spatiotemporal MRF approach to video segmentation: application to motion detection and lip segmentation, Signal Processing, 1999, 76(1): 6180. .[6]Park Sang Ho, Yun Dong, Lee Sang U K. Color image segmentation based on 3-D clustering: Morphological approach. Pattern Recognition, 1998, 31(8): 10611076. -
計(jì)量
- 文章訪問數(shù): 2697
- HTML全文瀏覽量: 104
- PDF下載量: 3185
- 被引次數(shù): 0
下載:
