基于显著对象的贝叶斯多目标检测方法

A Bayesian Multi-objects Detection Method Based on Salient Objects

针对多目标图像检测存在的误检问题,结合低层特征和中层提示,提出了一个新的基于显著对象的贝叶斯框架下的多目标检测方法。该方法首先用上下文感知显著检测方法获取图像的低层特征信息,然后用Ncut图像分割取得图像的显著中层信息提示,即多目标的类别标签信息,根据低层和中层信息提示来计算先验显著图,最后使用贝叶斯方法计算获得图像的后验显著图。实验结果表明,该方法提高了显著对象检测精度,并且可以较好地解决多目标检测误检问题。

There are problems of false detection when detecting image with multi-objects.In this paper,we propose a new multi-objects detection method which is based on salient objects within the Bayesian framework.First,we get the low level features via Context-Aware saliency detection.Then,we obtain the middle level cue by Ncut image segmentation which is category label information of multi-objects.The prior saliency map is computed with respected to both low and middle level cues.Last,we use a Bayesian formula to calculate the posteriori saliency map.The experimental result shows that,our method can better solve the problem of false detection of multi-object with higher detection precision.