H
Haochen Wang
Researcher at Beihang University
Publications - 10
Citations - 272
Haochen Wang is an academic researcher from Beihang University. The author has contributed to research in topics: Segmentation & Computer science. The author has an hindex of 3, co-authored 10 publications receiving 45 citations. Previous affiliations of Haochen Wang include Alibaba Group.
Papers
More filters
Book ChapterDOI
Few-Shot Semantic Segmentation with Democratic Attention Networks
TL;DR: This paper introduces the democratized graph attention mechanism, which can activate more pixels on the object to establish a robust correspondence between support and query images, and proposes multi-scale guidance by designing a refinement fusion unit to fuse features from intermediate layers for the segmentation of the query image.
Proceedings ArticleDOI
SwiftNet: Real-time Video Object Segmentation
TL;DR: SwiftNet as discussed by the authors compresses spatiotemporal redundancy in matching-based VOS via Pixel-Adaptive Memory (PAM), which adaptively triggers memory updates on frames where objects display noteworthy inter-frame variations.
Proceedings ArticleDOI
Long-range Attention Network for Multi-View Stereo
TL;DR: Wang et al. as mentioned in this paper introduced a Long-range Attention Network (LANet) to selectively aggregate reference features to each position to capture the long-range interdependence across the entire space.
Proceedings ArticleDOI
Variational Prototype Inference for Few-Shot Semantic Segmentation
TL;DR: In this paper, a probabilistic latent variable model is proposed to infer the distribution of the prototype that is treated as the latent variable, and the optimization is formulated as a variational inference problem with an amortized inference network based on an auto-encoder.
Proceedings ArticleDOI
Decoupled IoU Regression for Object Detection
TL;DR: In this article, a decoupled IoU Regression (DIR) model is proposed to handle the inconsistency between the confidence for NMS and the real localization confidence seriously affects detection performance.