Vision-Reading-Group

Table of Contents

• Overview
• Research Papers
  • 3D Vision
    • 3D Generation
    • 3D Editing
    • 3D Grounding
    • 3D Inpainting
    • NERFs
  • Diffusion
  • Distillation
• Survey Papers
  • Foundation Models
  • 3D Vision
    • NERFs
    • Emboidied AI
  • Adversarial
    • Data Poisoning
• Explainers
  • Diffusion

Overview

Mission

• Foster collaboration and knowledge exchange in various sub-fields of Computer Vision.
• Explore and discuss cutting-edge research (SOTA) across different areas.
• Provide a platform for brainstorming, peer review, and constructive criticism.
• Develop reading papers & ideation as a habit or hobby.

Format

• Online Sessions : Convenient for everyone.
• Semi-Formal : Develop reading and discussion skills in a relaxed atmosphere.
• Frequency : Weekly (optional for members to present every week).

Paper Selection

• Two Papers per Session : Focus on foundational or influential papers relevant to different sub-fields.
• Open Selection : Members can add interesting papers to a shared list (presenter not required to be the recommender).

Presentation

• Collaborative Exploration : Brief on-the-spot reading and group discussion encouraged. Members are free to prepare beforehand as well.
• Active Participation : Attendees should guide the presenter and contribute to understanding, not just ask questions.
• Ideas Sandbox : Members can present their own ideas for feedback and discussion.

Additional Points

• Minimal Commitment : While regular participation is encouraged, a minimum of two volunteers per week for paper presentations is preferred.
• Constructive Environment : Promote open discussion, respectful debate, and supportive feedback.

Research Papers

3D Vision

Feature Field Distillation

• Decomposing NeRF for Editing via Feature Field Distillation [Paper]
• Feature 3DGS: Supercharging 3D Gaussian Splatting to Enable Distilled Feature Fields [Paper]

3D Correspondence

• FlowMap: High-Quality Camera Poses, Intrinsics, and Depth via Gradient Descent [Submitted 23 Apr 2024 arxiv] [Paper] [Demo] ~Continue next time

3D Generation

• DreamFusion: Text-to-3D using 2D Diffusion [ICLR 23] [Paper] [Demo]
• LucidDreaming: Controllable Object-Centric 3D Generation [Submitted 30 Nov 2023 arxiv] [Paper] [Demo] [Code]

3D Editing

• Awesome NERF editing
• GaussianEditor: Swift and Controllable 3D Editing with Gaussian Splatting [Submitted 24 Nov 2023 arxiv] [Paper] [Demo] [Code] [Video]
• Instruct-NeRF2NeRF: Editing 3D Scenes with Instructions [ICCV 23] [Paper] [Demo] [Code]
• DreamBooth3D: Subject-Driven Text-to-3D Generation [ICCV 23] [Paper] [Demo]
• DreamEditor: Text-Driven 3D Scene Editing with Neural Fields [SIGGRAPH 23] [Paper] [Demo] [Code]

3D Grounding

• Language Conditioned Spatial Relation Reasoning for 3D Object Grounding [NerIPS 22] [Paper] [Demo] [Code]
• 3D Concept Grounding on Neural Fields [NeurIPS 22] [Paper] [Code]
• Multi-View Transformer for 3D Visual Grounding [CVPR 22] [Paper] [Code]

3D Inpainting

• Breathing New Life into 3D Assets with Generative Repainting [ICCV 23] [Paper] [Demo] [Code]
• RePaint: Inpainting using Denoising Diffusion Probabilistic Models [CVPR 22] [Paper] [Code]
• Diffusion Probabilistic Models for Scene-Scale 3D Categorical Data [Submitted 2 Jan 23 arxiv] [Paper] [Code]

NERFs

• NeRF: Representing Scenes as Neural Radiance Fields for View Synthesis [Paper] [Demo] [Code] [Dataset]
• D-NeRF: Neural Radiance Fields for Dynamic Scenes [CVPR 21] [Paper] [Demo] [Code]
• Mip-NeRF: A Multiscale Representation for Anti-Aliasing Neural Radiance Fields [ICCV 21] [Paper] [Code]

Diffusion

• Understanding Diffusion Models: A Unified Perspective. by Calvin Luo [Explainer, arxiv] [Paper]
• LatentPaint: Image Inpainting in Latent Space with Diffusion Models [WACV 24] [Paper] [Video]
• DreamBooth: Fine Tuning Text-to-Image Diffusion Models for Subject-Driven Generation [CVPR 23] [Paper] [Demo] [Dataset] [Tutorial]

Distillation

• Dreaming to Distill: Data-free Knowledge Transfer via DeepInversion [CVPR 20] [Paper] [Code]

Survey Papers

Foundation Models

• A Comprehensive Survey on Pretrained Foundation Models: A History from BERT to ChatGPT [Submitted 18 Feb 23 arxiv] [Paper]

3D Vision

NERFs

• NeRF: Neural Radiance Field in 3D Vision, A Comprehensive Review [Submitted 30 Nov 23 arxiv] [Paper]

Emboidied AI

• A Survey of Embodied AI: From Simulators to Research Tasks [IEEE Txn 22] [Paper]

Adversarial

Data Poisoning

• Just How Toxic is Data Poisoning? A Unified Benchmark for Backdoor and Data Poisoning Attacks [PMLR '21] [Paper]

Explainers

Diffusion

Understanding Diffusion Models: A Unified Perspective [Submitted 25 Aug 23 arxiv] [Paper]