基于外区抑制与马尔可夫模型的轮廓检测

Contour Detection Based on Surround Inhibition and Markov Model

轮廓检测是目标识别的重要前提,为解决复杂纹理自然场景下目标轮廓难以有效检测的问题,该文提出一种新的计算方法。首先,利用旋转不变滤波器快速实现外区抑制;其次,通过集合运算整合不同抑制水平的抑制后响应,得到边缘图;最后,建立基于马尔可夫随机场的轮廓概率模型,赋予每个边缘点一个概率值,选择概率较高的边缘点即得到最终的轮廓输出。定性和定量分析表明,相对于现有算法,新算法的轮廓检测性能显著提升,并具有更好的鲁棒性。

Contour detection plays an important role in object recognition. In order to detect effectively the target contour with complex textures in natural scenes, a new method of contour detection is proposed in this paper. In the algorithm, the surround inhibition is realized fastly by steerable filters; the multi-inhibited responses are integrated to form an edge map under the rules of set operations, and finally, according to the theory of Markov random field, a model of contour probability is presented to give each edge point a probability and form the output of contour by thresholding the probability map. Qualitative and quantitative analysis shows that the proposed method increases the performance signifacantly and get a better result of robustness, compared with existing methods.