HUST Vision Lab

[NeurIPS 2023] CircuitFormer: Circuit as Set of Points

hustvl

Lane Graph as Path: Continuity-preserving Path-wise Modeling for Online Lane Graph Construction

hustvl

Matte Anything: Interactive Natural Image Matting with Segment Anything Models

hustvl

hustvl

[CVPR 2023] PD-Quant: Post-Training Quantization Based on Prediction Difference Metric

hustvl

Query6DoF: Learning Sparse Queries as Implicit Shape Prior for Category-Level 6DoF Pose Estimation

hustvl

[CVPR 2023] RILS: Masked Visual Reconstruction in Language Semantic Space (https://arxiv.org/abs/2301.06958)

hustvl

hustvl

[Information Fusion] Boosting Image Matting with Pretrained Plain Vision Transformers

hustvl

A general map auto annotation framework based on MapTR, with high flexibility in terms of spatial scale and element type

hustvl