For natural images of complex texture, a supervised boundary detection method using the multi-information fusion was proposed.
针对自然图像纹理复杂的特点，提出了一种多种信息融合的有监督边界检测方法。
In order to enhance restriction for the synthesizing procedure, the paper synthesizes texture images by spiral patch, avoiding the discontinuous area happening in the synthesized image.
为加强合成过程的约束，采用螺旋顺序逐块合成纹理图像，避免了合成过程中不连续区域的出现。
The experimental results demonstrate the effectiveness of ISBEC for color-texture segmentation of natural images.
实验结果验证了ISBEC对自然图像色彩—纹理分割的有效性。
应用推荐