📚 딥러닝 논문읽기 스터디

• 매주 본인이 읽고싶은 논문을 하나씩 리뷰하고 다른 사람들 앞에서 발표하는 스터디입니다.
• 발표는 매주 목요일 오후 8시에 디스코드에서 진행하고 있습니다.

Image Enhancement (이미지 개선)

Title	Original Paper Link	Paper Review	reviewer
HDRUNet: Single Image HDR Reconstruction with Denoising and Dequantization	Link	Link	윤도현
Replacing Mobile Camera ISP with a Single Deep Learning Model	Link	Link	윤도현
RawHDR: High Dynamic Range Image Reconstruction from a Single Raw Image	Link	Link	윤도현

Object Detection & Segmentation (객체 탐지 & 분할)

Title	Original Paper Link	Paper Review	reviewer
DETRs with Collaborative Hybrid Assignments Training	Link	Link	윤도현

Depth Estimation (깊이 측정)

Title	Original Paper Link	Paper Review	reviewer
HITNet: Hierarchical Iterative Tile Refinement Network for Real-time Stereo Matching	Link	Link	윤도현
UniDepth: Universal Monocular Metric Depth Estimation	Link	Link	윤도현

3D Reconstruction (3차원 복원)

Title	Original Paper Link	Paper Review	reviewer
NeRF: Representing Scenes as Neural Radiance Fields for View Synthesis	Link	Link	윤도현
3D Gaussian Splatting for Real-Time Radiance Field Rendering	Link	Link	이하정
Scaffold-GS: Structured 3D Gaussians for View-Adaptive Rendering	Link	Link	이하정

Autonomous Driving (자율 주행)

Title	Original Paper Link	Paper Review	reviewer
DETR3D: 3D Object Detection from Multi-view Images via 3D-to-2D Queries	Link	Link	윤도현
Multi-Modal Fusion Transformer for End-to-End Autonomous Driving	Link	Link	이하정
BEVFusion: Multi-Task Multi-Sensor Fusion with Unified Bird’s-Eye View Representation	Link	Link	이태훈

LLM & Multi-modal (자연어처리 & 멀티모달)

Title	Original Paper Link	Paper Review	reviewer
Visual Instruction Tuning	Link	Link	이하정
LLaMA: Open and Efficient Foundation Language Models	Link	Link	이하정
Text2Loc: 3D Point Cloud Localization from Natural Language	Link	Link	이하정
Attention is All you Need	Link	Link	윤도현

Model Compression (모델 경량화)

Title	Original Paper Link	Paper Review	reviewer
Integer Quantization for Deep Learning Inference: Principles and Empirical Evaluation	Link	Link	윤도현

Computer Graphics

Title	Original Paper Link	Paper Review	reviewer
Drag Your GAN: Interactive Point-based Manipulation on the Generative Image Manifold	Link	Link	이하정
4D Gaussian Splatting for Real-Time Dynamic Scene Rendering	Link	Link	이하정

etc (기타)

Title	Original Paper Link	Paper Review	reviewer
DREAMFUSION: TEXT-TO-3D USING 2D DIFFUSION	Link	Link	이하정
PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation	Link	Link	이태훈
Swin Transformer: Hierarchical Vision Transformer using Shifted Windows	Link	Link	이태훈
Grounding Image Matching in 3D with MASt3R	Link	Link	이하정
ProlificDreamer: High-Fidelity and Diverse Text-to-3D Generation with Variational Score Distillation	Link	Link	이하정
FaceNet: A Unified Embedding for Face Recognition and Clustering	Link	Link	윤도현
Tracking Persons-of-Interest via Unsupervised Representation Adaptation	Link	Link	윤도현
Deep High-Resolution Representation Learning for Human Pose Estimation	Link	Link	윤도현

✅ 약속

• 다음 발표자는 일요일 밤 23:59까지 본인이 리뷰할 논문을 Issue에 등록해주세요.
• 발표가 끝나면 "본인의 깃허브 브랜치"에서 커밋한 뒤 당일 23:59까지 발표자료의 PR을 생성해주세요.

📋 커밋규칙

1. 저장소 local로 clone (최초 1번만)
$ git clone https://github.com/dohyeonYoon/Deep-Learning-Paper-Review.git

2. 먼저 본인의 로컬 메인 브랜치를 최신 상태로 업데이트 합니다.
$ git checkout main #메인 브랜치로 이동
$ git pull origin main #원격 저장소에서 최신 메인 브랜치 내용 받아오기
$ git checkout feature/ydh #작업 브랜치로 이동
$ git merge main #메인 브랜치 내용을 작업 브랜치에 병합

3. readme file을 수정하고 본인의 브랜치로 커밋
$ git add README.md
$ git commit -m "docs: 논문리뷰 추가, close #1" #본인의 이슈를 연동하고 PR이 병합되면 이슈닫힘
$ git push origin feature/ydh #작업 브랜치에 푸시

4. 깃허브에서 PR 생성

👨🏼‍💻 Members

윤도현	이하정	이태훈
Image Enhancement 3D Computer Vision	Computer Vision Computer Graphics Multi-modal	Automous Driving