Salient Object Detection Driven by Fixation Prediction
Salient Object Detection Driven by Fixation Prediction
复制标题
DOI：
10.1109/cvpr.2018.00184
复制
发表时间：
2018-06
期刊：
影响因子：
--
通讯作者：
Wenguan Wang;Jianbing Shen;Xingping Dong;A. Borji
中科院分区：
文献类型：
--
作者：
Wenguan Wang;Jianbing Shen;Xingping Dong;A. Borji
视觉显著性研究主要集中在两类模型上，即注视点预测和显著目标检测。然而，两者之间的关系却很少被探讨。在本文中，我们建议采用前模型类型来识别和分割场景中的显著对象。我们构建了一个名为Attentive Saliency Network（ASNet）1的新型神经网络，它可以学习从固定图中检测显著对象。在上层网络层导出的固定图捕获对场景的高级理解。然后，显著对象检测被视为细粒度的对象级显著性分割，并在固定图的指导下以自上而下的方式逐步优化。ASNet基于卷积LSTM（convLSTM）的层次结构，为分割图的顺序细化提供了一种有效的递归机制。为了提高ASNet的性能，引入了几个损失函数。广泛的实验评估表明，我们提出的ASNet是能够生成准确的分割图的帮助下计算的固定图。我们的工作提供了一个更深入的了解注意的机制和缩小之间的差距差距显著对象检测和固定预测。
Research in visual saliency has been focused on two major types of models namely fixation prediction and salient object detection. The relationship between the two, however, has been less explored. In this paper, we propose to employ the former model type to identify and segment salient objects in scenes. We build a novel neural network called Attentive Saliency Network (ASNet)1 that learns to detect salient objects from fixation maps. The fixation map, derived at the upper network layers, captures a high-level understanding of the scene. Salient object detection is then viewed as fine-grained object-level saliency segmentation and is progressively optimized with the guidance of the fixation map in a top-down manner. ASNet is based on a hierarchy of convolutional LSTMs (convLSTMs) that offers an efficient recurrent mechanism for sequential refinement of the segmentation map. Several loss functions are introduced for boosting the performance of the ASNet. Extensive experimental evaluation shows that our proposed ASNet is capable of generating accurate segmentation maps with the help of the computed fixation map. Our work offers a deeper insight into the mechanisms of attention and narrows the gap between salient object detection and fixation prediction.