Object Detection for Remote Sensing Image Based on Top-Down Visual Attention
Tao Fang · Jisuanji gongcheng · 2011
A novel method based on Top-Down visual attention is proposed for object detection in remote sensing image.The method considers both object appearance cue and context cue in Top-Down attention.Object appearance cue is used to choose regions with the same features as object appearance to improve the efficiency.Context cue is used to focus the search on object context environment to confirm the correctness.Experimental results validate the good performance of the method in improving both the efficiency and the accuracy of object detection.