Jiacheng Chen
Simon Fraser University
16 Papers
23 Citations
Jiacheng Chen is an academic researcher from Simon Fraser University. The author has contributed to research in topics: Computer science & Rendering (computer graphics). The author has an hindex of 5, co-authored 8 publications.
Chat about Author
Papers
Learning the Best Pooling Strategy for Visual Semantic Embedding
Jiacheng Chen,Hexiang Hu,Hao Wu,Yuning Jiang,Changhu Wang +4 more
- 01 Jun 2021
TL;DR: Wang et al. as discussed by the authors proposed a generalized pooling operator (GPO) to automatically adapt itself to the best pooling strategy for different features, requiring no manual tuning while staying effective and efficient.
•Posted Content
Learning the Best Pooling Strategy for Visual Semantic Embedding.
TL;DR: A Generalized Pooling Operator (GPO) is proposed, which learns to automatically adapt itself to the best pooling strategy for different features, requiring no manual tuning while staying effective and efficient and can be a plug-and-play feature aggregation module for standard VSE models.
MVDiffusion: Enabling Holistic Multi-view Image Generation with Correspondence-Aware Diffusion
TL;DR: MVDiffusion as mentioned in this paper proposes a correspondence-aware attention mechanism, enabling effective cross-view interaction, and demonstrates the first method capable of generating a textured map of a scene mesh.
84
Improving Pixel-based MIM by Reducing Wasted Modeling Capability
Yuan Liu,Songyang Zhang,Jiacheng Chen,Zhaohui Yu,Kai Chen,Dahua Lin +5 more
TL;DR: This paper is the first to systematically investigate multilevel feature fusion for isotropic architectures like the standard Vision Transformer (ViT) and is the first to systematically investigate multilevel feature fusion for isotropic architectures like the standard Vision Transformer (ViT).
•Proceedings Article
Probabilistic Neural Programmed Networks for Scene Generation
Zhiwei Deng,Jiacheng Chen,Yifang Fu,Greg Mori +3 more
- 01 Jan 2018
TL;DR: PNP-Net is proposed, a variational auto-encoder framework that flexibly composes images with a dynamic network structure, learns a set of distribution transformers that can compose distributions based on semantics, and decodes samples from these distributions into realistic images.