Edge-assisted Object Segmentation using Multimodal Feature Aggregation and Learning

Object segmentation aims to perfectly identify objects embedded in the surrounding environment and has a wide range of applications. Most previous methods object only use RGB images ignore geometric information from disparity images. Making full heterogeneous data different devices proved be very effective strategy for improving performance. The key challenge multimodal fusion based task lies learning, transformation, information. In this paper, we focus on transformation features. We develop framework, termed Hybrid Fusion Segmentation Network (HFSNet). Specifically, HFSNet contains three components, i.e., convolutional sparse coding (DCSC), asymmetric dense projection feature aggregation (ADPFA) (MFF). DCSC is designed coding. It not better interpretability but also preserves object. ADPFA enhance texture fully exploit nonadjacent MFF used perform fusion. Extensive experiments show that our outperforms existing state-of-the-art models two challenging datasets.