360VOTS: Visual Object Tracking and Segmentation in Omnidirectional Videos

Yinzhe Xu, Huajian Huang, Yingshu Chen, Sai-Kit Yeung
Arxiv 2024.

360VOTS is a benchmark dataset for omnidirectional object tracking and segmentation, comprising 360VOT and 360VOS datasets. 360VOT offers unbiased ground truth of (rotated) bounding boxes and field-of-views, while 360VOS provides pixel-wise masks. In addition, a general 360 tracking framework and new metrics are tailored for both omnidirectional visual object tracking and segmentation tasks.

StyleCity: Large-Scale 3D Urban Scenes Stylization

Yingshu Chen, Huajian Huang, Tuan-Anh Vu, Ka Chun Shum, Sai-Kit Yeung
European Conference on Computer Vision (ECCV) 2024.

StyleCity is a vision-and-text-driven texture stylization system for large-scale 3D urban scenes. It synthesizes style-aligned urban texture and 360° sky background, while keeping the scene identity intact.

Advances in 3D Neural Stylization: A Survey

Yingshu Chen, Guocheng Shao, Ka Chun Shum, Binh-Son Hua, Sai-Kit Yeung
Arxiv 2023.

Comprehensive review for neural stylization on 3D data, including meshes, neural fields, point clouds, volume, implicit shapes.

360VOT: A New Benchmark Dataset for Omnidirectional Visual Object Tracking

Huajian Huang, Yinzhe Xu, Yingshu Chen, Sai-Kit Yeung
International Conference on Computer Vision (ICCV) 2023.

360VOT aims at the novel task of omnidirectional object tracking with new video object tracking benchmark dataset and evaluation metrics.

360Roam: Real-Time Indoor Roaming Using Geometry-Aware 360ᵒ Radiance Fields

Huajian Huang, Yingshu Chen, Tianjian Zhang, Sai-Kit Yeung
Arxiv 2022 | SIGGRAPH Asia 2022 Technical Communications.

360Roam is a novel scene-level NeRF system that can synthesize novel views of large-scale multi-room indoor scenes and support real-time indoor roaming.

Neural Scene Decoration from a Single Photograph

Hong Wing Pang, Yingshu Chen, Phuoc-Hieu T. Le, Binh-Son Hua, Thanh Nguyen, Sai-Kit Yeung
European Conference on Computer Vision (ECCV) 2022.

Given a photograph of an empty indoor space and a list of decorations with layout determined by user, we aim to synthesize a new image of the same space with desired furnishing and decorations.

Time-of-Day Neural Style Transfer for Architectural Photographs

Yingshu Chen, Tuan-Anh Vu, Ka-Chun Shum, Binh-Son Hua, Sai-Kit Yeung
International Conference on Computational Photography (ICCP) 2022.

Architectural Photography Style Transfer aims to transfer background dynamic texture and chrominance, and transfer sufficient styles for foreground while keeping foreground geometry intact.