2022.10.11@更新两篇论文的阅读
2022.10.15@更新一篇UGTR的论文阅读
2022.12.03@更新不知道几篇

伪装目标检测入门

(Arxiv2023)Forgery-aware Adaptive Transformer for Generalizable Synthetic Image Detection  

篡改图像检测论文 - Part2

篡改图像检测论文 - Part1

伪装目标检测论文 - Part1

Domain Generalization - Part1

Diffusion Model & AutoEncoder

Some MOE of Image Classification

Training CNNs in Presence of JPEG Compression: Multimedia Forensics vs Computer Vision 论文阅读

ARE GAN GENERATED IMAGES EASY TO DETECT? A CRITICAL ANALYSIS OF THE STATE-OF-THE-ART 论文阅读

【V-MoE】论文学习笔记

ClassSR: A General Framework to Accelerate Super-Resolution Networks by Data Characteristic 论文阅读

【深度】动态神经网络综述

DM检测论文阅读

Derivation of formulas in Score-Based Models

(CVPR2023)DCFace: Synthetic Face Generation with Dual Condition Diffusion Model 论文阅读

RGBD2: Generative Scene Synthesis via Incremental View Inpainting using RGBD Diffusion Models 论文阅读

Prompt-to-Prompt Image Editing with Cross Attention Control 论文阅读

Magic3D: High-Resolution Text-to-3D Content Creation 论文阅读

(AAAI2022-PSOD)Weakly-Supervised Salient Object Detection Using Point Supervision 论文阅读

(CVPR2023)DiffRF: Rendering-Guided 3D Radiance Field Diffusion 论文阅读

(CVPR2023)Diffusion-SDF: Text-to-Shape via Voxelized Diffusion

(ICLR2023)DIFFUSION PROBABILISTIC FIELDS

CAL：低分辨率姿态估计

(CVPR2023-IOCFormer)Indiscernible Object Counting in Underwater Scenes 论文阅读

1. 扩散过程的最终结果  不是高斯噪声, 而是 Coarse Estimate Distribution. 
2. 使用 GMM 重新建模 , 使得扩散的中间结果的分布带有方差信息.

@2023.06.03 突然发现这篇文章中CVPR了 质量确实不错

(CVPR2023)DiffPose: Toward More Reliable 3D Pose Estimation 论文阅读

(ECCV2022)UIA-ViT: Unsupervised Inconsistency-Aware Method based on Vision Transformer for Face Forgery Detection 论文阅读

(ICML2023)NerfDiff: Single-image View Synthesis with NeRF-guided Distillation from 3D-aware Diffusion 论文阅读

Weakly-Supervised Camouflaged Object Detection with Scribble Annotations 论文阅读

(AAAI2021-SCWSSOD)Structure-Consistent Weakly Supervised Salient Object Detection with Local Saliency Coherence 论文阅读

感觉DGNet就是从里面拿了一部分

(CVPR2020-ScribbleSaliency)Weakly-Supervised Salient Object Detection via Scribble Annotations

初看感觉没道理 再看其实挺有道理的 感觉可以直接用到弱监督

(CVPR2023)HOLODIFFUSION: Training a 3D Diffusion Model using 2D Images-论文阅读

2022.10.15@更新了一点, 发现还是得看的 QAQ
2022.10.20@继续更新了一点
2022.10.26@添加了LDM
2023.05.06@修改了一处公式错误

Diffusion Models - 扩散模型

2023.05.23@更加细致的读了一下
2023.05.26@修改了一处公式错误

(Song)Consistency Model 论文阅读

DiffusionSeg: Adapting Diffusion Towards Unsupervised Object Discovery 论文阅读

(CVPR2023) DiffusionRig: Learning Personalized Priors for Facial Appearance Editing 论文阅读

BerDiff: Conditional Bernoulli Diffusion Model for Medical Image Segmentation 论文阅读

(CVPR2023)Ambiguous Medical Image Segmentation using Diffusion Models 论文阅读

(CVPR2023-FEDER)Camouflaged Object Detection with Feature Decomposition and Edge Reconstruction 论文阅读

新的 Baseline 就是新的方向可以做 还是多模态的

(CVPR2023-DGM)Detecting and Grounding Multi-Modal Media Manipulation 论文阅读

(WACV2022-OCENet)Modeling Aleatoric Uncertainty for Camouflaged Object Detection 论文简读

Improving Camouflaged Object Detection with the Uncertainty of Pseudo-edge Labels 论文阅读

Cold Diffusion: Inverting Arbitrary Image Transforms Without Noise

Few-shot learning（少样本学习） 

MaskDiff: Modeling Mask Distribution with Diffusion Probabilistic Model for Few-Shot Instance Segmentation 论文阅读

SegFormer语义分割

A CONDITIONAL POINT DIFFUSION-REFINEMENT PARADIGM FOR 3D POINT CLOUD COMPLETION 论文阅读

Stepwise Feature Fusion: Local Guides Global 论文阅读

All are Worth Words: A ViT Backbone for Diffusion 论文阅读

TransUNet 论文阅读

Feature Shrinkage Pyramid for Camouflaged Object Detection with Transformers 论文阅读

总结了两篇文章，添加了一些注解

DETR 详解

Boosting Camouflaged Object Detection with Dual-Task Interactive Transformer 论文阅读

A Generalist Framework for Panoptic Segmentation of Images and Videos 论文阅读

Activate Learning 入门

End-to-End Reconstruction-Classification Learning for Face Forgery Detection 论文阅读

15 分钟搞定 Softmax Loss 求导

详解两阶段 3D 目标检测网络 Voxel R-CNN：Towards High Performance Voxel-based 3D Object Detection

SegFix论文阅读

Global average pooling 与 everage pooling 的区别

Global average pooling (GAP)

PointNet 系列论文解读

有部分参考自知乎, 链接已给出

PV-RCNN 阅读笔记

3W 字长文带你轻松入门视觉 transformer

3D 目标检测深度学习方法之 voxel-represetnation 内容综述（一）

Deformable Convolution  v1, v2 总结

SECOND 算法

这个是我参加MAC实验室考核的时候报告的小结, 整理了一下发布出来

Sparse R-CNN阅读笔记

【3D 目标检测】PointPillars 论文和代码解析

扔掉 anchor！真正的 CenterNet——Objects as Points 论文解读

Waymo 新作: 端到端多角度融合的 3D 点云目标检测

Comprehensive Multi-Modal Interactions for Referring Image Segmentation 论文阅读笔记

Self-training with Noisy Student

PVT(Pyramid Vision Transformer)

OHEM 详细解析

CV+Transformer 之 Swin Transformer

实例分割新思路之 SOLO v1&v2 深度解析

Swin Transformer V2 论文解析

torch.utils.data.distributed.DistributedSampler 解析

pytorch 混合精度训练

也来盘点一些最近的非 Transformer 工作

Dynamic Sparse R-CNN

Pytorch 分布式训练

GAN 评价指标

Synthesizer: Rethinking Self-Attention in Transformer Models

相对位置编码 Transformer 的一个理论缺陷与对策

实例分割的进阶三级跳：从 Mask R-CNN 到 Hybrid Task Cascade

Deformable DETR

语义分割损失函数合集

Anchor DETR

一文读懂 Faster RCNN

总是忘记 ROI Align 和 ROI Pooling的区别，还是复制以下吧

ROI Pooling 与 ROI Align

这篇文章开头提出目前的伪造检测算法，在训练集与测试集都是使用同一个伪造生成算法的情况下，检测效果很好。（即普通DeepFakes训练过程中的判别器，其输入的假图像分布一致，都是由同一个生成器生成的）这篇文章使用的GAN结构与DeepFakes不同，其使用设计的算法生成不同方式伪造的假图像，再由判别器判断是否为假图像。
为了提高伪造效果，使用对抗方法来生成判别器鉴别困难的伪造配置。另外，引入多任务学习来提高学习效率，判别器还要判断输入假图像的伪造配置。

Self-supervised Learning of Adversarial Example: Towards Good Generalizations for Deepfake Detection

Non-Local neural networks 的理解与实现

Advances, challenges and opportunities in creating data for trustworthy AI 笔记

UP-DETR: Unsupervised Pre-training for Object Detection with Transformers 论文阅读

SegNeXt: Rethinking Convolutional Attention Design for Semantic Segmentation 笔记

Uncertainty Modeling for Out-of-Distribution Generalization 论文阅读

CE-Net: Context Encoder Network for 2D Medical Image Segmentation 论文阅读

ILVR: Conditioning Method for Denoising Diffusion Probabilistic Models 论文笔记

DALL·E 2 解读 | 结合预训练 CLIP 和扩散模型实现文本 - 图像生成