scene-analysis-and-understanding.md

CVPR-2023-Papers

Application
New collections

Scene Analysis and Understanding

Title	Repo	Paper	Video
You Only Segment Once: Towards Real-Time Panoptic Segmentation
IS-GGT: Iterative Scene Graph Generation with Generative Transformers
Disentangling Orthogonal Planes for Indoor Panoramic Room Layout Estimation with Cross-Scale Distortion Awareness			➖
Panoptic Video Scene Graph Generation
3D Spatial Multimodal Knowledge Accumulation for Scene Graph Prediction in Point Cloud	➖
JacobiNeRF: NeRF Shaping with Mutual Information Gradients
Learning Geometric-Aware Properties in 2D Representation using Lightweight CAD Models, or Zero Real 3D Pairs
Learning and Aggregating Lane Graphs for Urban Automated Driving
MIME: Human-Aware 3D Scene Generation			➖
Connecting the Dots: Floorplan Reconstruction using Two-Level Queries
NeRF-RPN: A General Framework for Object Detection in NeRFs
Relational Context Learning for Human-Object Interaction Detection
Symmetric Shape-Preserving Autoencoder for Unsupervised Real Scene Point Cloud Completion
Token Contrast for Weakly-Supervised Semantic Segmentation			➖
MM-3DScene: 3D Scene Understanding by Customizing Masked Modeling with Informative-Preserved Reconstruction and Self-Distilled Consistency
Primitive Generation and Semantic-related Alignment for Universal Zero-Shot Segmentation
CLIP2Scene: Towards Label-Efficient 3D Scene Understanding by CLIP
Multispectral Video Semantic Segmentation: A Benchmark Dataset and Baseline			➖
Optimal Transport Minimization: Crowd Localization on Density Maps for Semi-Supervised Counting
Indiscernible Object Counting in Underwater Scenes
Long Range Pooling for 3D Large-Scale Scene Understanding			➖
Delivering Arbitrary-Modal Semantic Segmentation
Images Speak in Images: A Generalist Painter for In-Context Visual Learning			➖
SCPNet: Semantic Scene Completion on Point Cloud
Content-Aware Token Sharing for Efficient Semantic Segmentation with Vision Transformers
OpenScene: 3D Scene Understanding with Open Vocabularies
Devil's on the Edges: Selective Quad Attention for Scene Graph Generation
Delving into Shape-Aware Zero-Shot Semantic Segmentation
Category Query Learning for Human-Object Interaction Classification
Nerflets: Local Radiance Fields for Efficient Structure-Aware 3D Scene Representation from 2D Supervision
DejaVu: Conditional Regenerative Learning to Enhance Dense Prediction	➖		➖
SCOOP: Self-Supervised Correspondence and Optimization-based Scene Flow
Incremental 3D Semantic Scene Graph Prediction from RGB Sequences
PanelNet: Understanding 360 Indoor Environment via Panel Representation	➖
Perspective Fields for Single Image Camera Calibration
Open-Category Human-Object Interaction Pre-Training via Language Modeling Framework	➖		➖
Fast Contextual Scene Graph Generation with Unbiased Context Augmentation			➖
Diffusion-based Generation, Optimization, and Planning in 3D Scenes
TopNet: Transformer-based Object Placement Network for Image Compositing	➖
Computational Flash Photography through Intrinsics
Probing Neural Representations of Scene Perception in a Hippocampally Dependent Task using Artificial Neural Networks	➖		➖
DeepSolo: Let Transformer Decoder with Explicit Points Solo for Text Spotting			➖
LEGO-Net: Learning Regular Rearrangements of Objects in Rooms
Open-Vocabulary Point-Cloud Object Detection without 3D Annotation
Weakly-Supervised Domain Adaptive Semantic Segmentation with Prototypical Contrastive Learning
ScanDMM: A Deep Markov Model of Scanpath Prediction for 360° Images
Canonical Fields: Self-Supervised Learning of Pose-Canonicalized Neural Fields
TempSAL - Uncovering Temporal Information for Deep Saliency Prediction
Probabilistic Debiasing of Scene Graphs
Towards Unified Scene Text Spotting based on Sequence Generation			➖
Learning to Generate Language-Supervised and Open-Vocabulary Scene Graph using Pre-trained Visual-Semantic Space
Modular Memorability: Tiered Representations for Video Memorability Prediction
Where we are and what we're Looking at: Query based Worldwide Image Geo-Localization using Hierarchies and Scenes
HRDFuse: Monocular 360° Depth Estimation by Collaboratively Learning Holistic-with-Regional Depth Distributions