Litcius/Paper detail

CASNet: A Cross-Attention Siamese Network for Video Salient Object Detection

Yuzhu Ji, Haijun Zhang, Zequn Jie, Lin Ma, Q. M. Jonathan Wu

2020IEEE Transactions on Neural Networks and Learning Systems119 citationsDOI

Abstract

Recent works on video salient object detection have demonstrated that directly transferring the generalization ability of image-based models to video data without modeling spatial-temporal information remains nontrivial and challenging. Considering both intraframe accuracy and interframe consistency of saliency detection, this article presents a novel cross-attention based encoder-decoder model under the Siamese framework (CASNet) for video salient object detection. A baseline encoder-decoder model trained with Lovász softmax loss function is adopted as a backbone network to guarantee the accuracy of intraframe salient object detection. Self- and cross-attention modules are incorporated into our model in order to preserve the saliency correlation and improve intraframe salient detection consistency. Extensive experimental results obtained by ablation analysis and cross-data set validation demonstrate the effectiveness of our proposed method. Quantitative results indicate that our CASNet model outperforms 19 state-of-the-art image- and video-based methods on six benchmark data sets.

Topics & Concepts

Computer scienceBenchmark (surveying)Artificial intelligenceEncoderSalientConsistency (knowledge bases)Computer visionPattern recognition (psychology)GeneralizationInter frameObject detectionFrame (networking)Reference frameMathematicsTelecommunicationsGeographyOperating systemMathematical analysisGeodesyVisual Attention and Saliency DetectionImage and Video Quality AssessmentFace Recognition and Perception
CASNet: A Cross-Attention Siamese Network for Video Salient Object Detection | Litcius