Yanan Sun

papers

1,595

total citations

papers (20)

GSNet: Joint Vehicle Pose and Shape Reconstruction with Geometrical and Scene-aware Supervision

ECCV 2020arXiv

citations

A Unified Query-Based Paradigm for Point Cloud Understanding

CVPR 2022arXiv

citations

Human Instance Matting via Mutual Guidance and Multi-Instance Refinement

CVPR 2022arXiv

citations

Homogeneous Architecture Augmentation for Neural Predictor

ICCV 2021arXiv

citations

FaceShot: Bring Any Character into Life

ICLR 2025arXiv

citations

Towards Accurate and Robust Architectures via Neural Architecture Search

CVPR 2024arXiv

citations

Communication-efficient Federated Learning with Single-Step Synthetic Features Compressor for Faster Convergence

ICCV 2023arXiv

citations

Zero-cost Proxy for Adversarial Robustness Evaluation

ICLR 2025

citations

Ultrahigh Resolution Image/Video Matting With Spatio-Temporal Sparsity

CVPR 2023

citations

CARL: Causality-guided Architecture Representation Learning for an Interpretable Performance Predictor

ICCV 2025arXiv

citations

CN: Channel Normalization For Point Cloud Recognition

ECCV 2020

citations

Autoregressive Stylized Motion Synthesis With Generative Flow

CVPR 2021

citations

Audio-Driven Stylized Gesture Generation with Flow-Based Model

ECCV 2022

citations

Bridge the Gap Between Architecture Spaces via A Cross-Domain Predictor

NEURIPS 2022

citations

Loss Functions for Predictor-based Neural Architecture Search

ICCV 2025arXiv

citations

Position: Exploring the Robustness of Pipeline-Parallelism-Based Decentralized Training

ICML 2024

citations

Yanan Sun

papers (20)

3DSSD: Point-Based 3D Single Stage Object Detector

Semantic Image Matting

Visual Point Cloud Forecasting enables Scalable Autonomous Driving

Deep Video Matting via Spatio-Temporal Alignment and Aggregation

GSNet: Joint Vehicle Pose and Shape Reconstruction with Geometrical and Scene-aware Supervision

A Unified Query-Based Paradigm for Point Cloud Understanding

Human Instance Matting via Mutual Guidance and Multi-Instance Refinement

Homogeneous Architecture Augmentation for Neural Predictor

FaceShot: Bring Any Character into Life

Towards Accurate and Robust Architectures via Neural Architecture Search

Communication-efficient Federated Learning with Single-Step Synthetic Features Compressor for Faster Convergence

Zero-cost Proxy for Adversarial Robustness Evaluation

Ultrahigh Resolution Image/Video Matting With Spatio-Temporal Sparsity

CARL: Causality-guided Architecture Representation Learning for an Interpretable Performance Predictor

CN: Channel Normalization For Point Cloud Recognition

Autoregressive Stylized Motion Synthesis With Generative Flow

Audio-Driven Stylized Gesture Generation with Flow-Based Model

Bridge the Gap Between Architecture Spaces via A Cross-Domain Predictor

Loss Functions for Predictor-based Neural Architecture Search

Position: Exploring the Robustness of Pipeline-Parallelism-Based Decentralized Training

papers (20)

3DSSD: Point-Based 3D Single Stage Object Detector

Semantic Image Matting

Visual Point Cloud Forecasting enables Scalable Autonomous Driving

Deep Video Matting via Spatio-Temporal Alignment and Aggregation

GSNet: Joint Vehicle Pose and Shape Reconstruction with Geometrical and Scene-aware Supervision

A Unified Query-Based Paradigm for Point Cloud Understanding

Human Instance Matting via Mutual Guidance and Multi-Instance Refinement

Homogeneous Architecture Augmentation for Neural Predictor

FaceShot: Bring Any Character into Life

Towards Accurate and Robust Architectures via Neural Architecture Search

Communication-efficient Federated Learning with Single-Step Synthetic Features Compressor for Faster Convergence

Zero-cost Proxy for Adversarial Robustness Evaluation

Ultrahigh Resolution Image/Video Matting With Spatio-Temporal Sparsity

CARL: Causality-guided Architecture Representation Learning for an Interpretable Performance Predictor

CN: Channel Normalization For Point Cloud Recognition

Autoregressive Stylized Motion Synthesis With Generative Flow

Audio-Driven Stylized Gesture Generation with Flow-Based Model

Bridge the Gap Between Architecture Spaces via A Cross-Domain Predictor

Loss Functions for Predictor-based Neural Architecture Search

Position: Exploring the Robustness of Pipeline-Parallelism-Based Decentralized Training