Acta Photonica Sinica, Volume. 50, Issue 6, 197(2021)
Combining MRF Energy and DCE-MRI Time-domain Features for Breast Tumors Segmentation Algorithm
To solve the low contrast, blurred boundry and intensity inhomogeous of the breast cancer lesions in the dynamic contrast-enhanced magnetic resonance imaging images, an integrated active contour model is proposed by combining markov random field energy with time-domain features. First, the edge-stop function of active contour model is derived from a fuzzy c-means cluster which treat the intensity and variation of time-domain as the feature. Then, markov random field energy is constructed to improve the difference between the lesions and other tissues. Finally, the region term is derived from k-nearest neighbor method which treat markov random field energy as dataset. The evolution of the contour curve stops at the boundary of lesions, and the energy function constructed by region term and edge term is minimized.The experiment proved that markov random field energy and time-domain feature can improve the contrast between the breast tumours and other tissues. Compared with state of the art of active contours models, the result segmented by the proposed method is more similar to the artificial segmentation, so that the proposed method is meaningful for breast cancer segmentation.
Get Citation
Copy Citation Text
Haoyang ZHOU, Bao FENG, Feifei QI, Zhuangsheng LIU, Wansheng LONG. Combining MRF Energy and DCE-MRI Time-domain Features for Breast Tumors Segmentation Algorithm[J]. Acta Photonica Sinica, 2021, 50(6): 197
Category: Image Processing
Received: Dec. 2, 2020
Accepted: Mar. 18, 2021
Published Online: Aug. 31, 2021
The Author Email: FENG Bao (fengbao1986.love@163.com)