目标检测基础

参考文章

22 款神经网络的设计和可视化工具【转载】

https://blog.csdn.net/zong596568821xp/article/details/107283252

github如何查看项目历史版本

https://www.jianshu.com/p/feef717fb52b

超越EfficientNet！MutualNet：一种自适应相互学习网络宽度和分辨率的网络

https://blog.csdn.net/weixin_47196664/article/details/107588014

PyTorch image models, scripts, pretrained weights -- (SE)ResNet/ResNeXT, DPN, EfficientNet, MixNet, MobileNet-V3/V2, MNASNet, Single-Path NAS, FBNet, and more

https://github.com/rwightman/pytorch-image-models

OpenMMLab Detection Toolbox and Benchmark，不断优化的目标检测系列项目

https://github.com/open-mmlab/mmdetection

PyImageSearch - You can master Computer Vision, Deep Learning, and OpenCV.——国外项目

https://www.pyimagesearch.com/

【AI不惑境】深度学习中的多尺度模型设计

https://www.jianshu.com/p/57cfa4fdd423

如何自己打造一个深度学习服务器？ - 知乎

https://zhuanlan.zhihu.com/p/40306143

The Open Images dataset，针对大场景，多标签，多目标

https://github.com/openimages/dataset

模型优化挂件，由YOLOv4启发

解读Squeeze-and-Excitation Networks（SENet）

https://www.jianshu.com/p/59fdc448a33f

Code for "Jointly Optimize Data Augmentation and Network Training: Adversarial Data Augmentation in Human Pose Estimation" (CVPR 2018)，对抗数据增强，YOLOv4中采用

https://github.com/zhiqiangdon/pose-adv-aug

目标检测特殊层：PSROIPooling详解

https://blog.csdn.net/wfei101/article/details/80766812

Focal Loss理解，RetinaNet中提出

https://www.cnblogs.com/king-lps/p/9497836.html

Focal loss论文详解

https://zhuanlan.zhihu.com/p/49981234

Deformable Convolution Networks[译]

https://www.jianshu.com/p/940d21c79aa3

【论文复现】CBAM: Convolutional Block Attention Module，YOLOv4中采用

https://blog.csdn.net/u013738531/article/details/82731257

实例分割--(PANet)Path Aggregation Network for Instance Segmentation，YOLOv4采用

https://blog.csdn.net/u011974639/article/details/79595179

增强CNN学习能力的Backbone:CSPNet，YOLOv4采用

https://blog.csdn.net/DD_PP_JJ/article/details/105094761

EfficientDet-BiFPN，YOLOv4采用

https://blog.csdn.net/qq_36808245/article/details/103603020

模型蒸馏

CNN channel pruning, LeGR, MorphNet, AMC. Codebase for paper "LeGR: Filter Pruning via Learned Global Ranking"，剪枝，求秩

https://github.com/cmu-enyac/LeGR

【DL】模型蒸馏Distillation

https://zhuanlan.zhihu.com/p/71986772

multi-teacher学习

https://blog.csdn.net/yagreenhand/article/details/104394008

剪枝

基于MobileNetV3的结构性剪枝优化 - 道客巴巴

http://www.doc88.com/p-13647389770455.html

网络压缩-2，剪枝，3、量化，4、降低数据数值范围 ，5、迁移学习

https://blog.csdn.net/J_Boom/article/details/87933448

清华&伯克利ICLR论文：重新思考6大剪枝方法

https://www.sohu.com/a/270464822_100024677

模型通道剪枝汇总（channel pruning），总结的好，有翻译，有代码，有论文，有工具

https://blog.csdn.net/qq_38109843/article/details/107462887

网络模型剪枝-论文阅读《Pruning Filters For Efficient Convnets》

https://blog.csdn.net/AManFromEarth/article/details/106888694

yolov3通道剪枝层剪枝知识蒸馏：无人机数据集案例

https://blog.csdn.net/weixin_41397123/article/details/103828931

论文阅读【模型剪枝（二）】

https://blog.csdn.net/weixin_38632246/article/details/95975928

目标追踪数据集

伯克利发布BDD100K：目前最大规模开放驾驶视频数据集

https://baijiahao.baidu.com/s?id=1602056615202328671&wfr=spider&for=pc

航空航天遥感图像目标检测数据集汇总

https://blog.csdn.net/weixin_36670529/article/details/84580431

最新的无人机视觉综述，数据集VisDrone现已开源

https://zhuanlan.zhihu.com/p/103582224

注意力机制

详解深度学习中“注意力机制”

https://www.jianshu.com/p/9b922fb83d77

注意力机制最新综述解读

https://blog.csdn.net/shenziheng1/article/details/89323074

空间域和通道域注意力机制

https://blog.csdn.net/zfnice/article/details/94998471

【论文复现】CBAM: Convolutional Block Attention Module

https://blog.csdn.net/u013738531/article/details/82731257

注意力机制CBAM代码实现（续篇）

https://blog.csdn.net/qq_43265072/article/details/106058693

多标签问题

分类问题-----多标签（multilabel）、多类别（multiclass）

https://blog.csdn.net/qq_27009517/article/details/80264919

深度学习---多标签分类问题

https://blog.csdn.net/qq_38906523/article/details/80210527

多类分类问题实现（np_utils.to_categorical函数使用）

https://www.cnblogs.com/lhuser/p/9073012.html

人体姿态估计，密集链接

卷积神经网络的网络结构——Hourglass

https://blog.csdn.net/u013841196/article/details/81048237

深度学习模块介绍 —— Hourglass Module

https://blog.csdn.net/BeBuBu/article/details/102935262

CenterNet的骨干网络之DLASeg

https://blog.csdn.net/DD_PP_JJ/article/details/107833023

Code for the CVPR Paper "Deep Layer Aggregation"——DLA网络

https://github.com/ucbdrive/dla

cvpr2018 Deep Layer Aggregation（DLANet）

https://blog.csdn.net/wuyubinbin/article/details/80622762

Deep Layer Aggregation笔记

https://blog.csdn.net/weeeeeida/article/details/88106665

CenterNet(Objects as points)代码解析，好读好用

https://github.com/zzzxxxttt/pytorch_simple_CenterNet_45

大神CSDN：https://blog.csdn.net/DD_PP_JJ

pytorch版CenterNet训练自己的数据集

https://blog.csdn.net/DD_PP_JJ/article/details/107715634

CenterNet 数据加载解析

https://blog.csdn.net/DD_PP_JJ/article/details/107691811

CenterNet骨干网络之hourglass

https://blog.csdn.net/DD_PP_JJ/article/details/107715399

CenterNet的骨干网络之DLASeg

https://blog.csdn.net/DD_PP_JJ/article/details/107833023

CenterNet之loss计算代码解析

https://blog.csdn.net/DD_PP_JJ/article/details/108065543

CenterNet测试推理过程

https://blog.csdn.net/DD_PP_JJ/article/details/108350727

SSD和RetinaFace的代码解析，棒！

Pytorch-RetinaFace 详解，biubug6/Pytorch_Retinaface

https://zhuanlan.zhihu.com/p/157297085

【SSD算法】史上最全代码解析-核心篇，基于ssd.pytorch

https://zhuanlan.zhihu.com/p/79854543

目标检测算法之NMS后处理相关

https://mp.weixin.qq.com/s/orYMdwZ1VwwIScPmIiq5iA

SSD-pytorch和RetinaFace-pytorch移植到NCNN和android上：

ssd.pytorch to ncnn，较符合pytorch版本的代码移植需求

https://github.com/xuming0629/ncnn_ssd

MobileNet-SSD Face Detection by ncnn，可参考模型导入结构

https://github.com/imistyrain/ncnn_face

mobilenet ssd @ ncnn，使用较早版本的ncnn，有一些参考意义

https://github.com/arlose/ncnn-mobilenet-ssd

using ncnn, on android, detection, model size only 11MB, 10fps on 845/cpu，结合ncnn和SSD移植到android

https://github.com/JuZiSYJ/MobilenetSSD_Android

Anchor-free网络：

Object detection, 3D detection, and pose estimation using center point detection:——开创性工作，好

https://github.com/xingyizhou/CenterNet

CenterNet的骨干网络之DLASeg

https://blog.csdn.net/DD_PP_JJ/article/details/107833023

tensorboy/centerpose: Push the Extreme of the pose estimation

https://github.com/tensorboy/centerpose

目标检测anchor free系列检测器简介：CornerNet,CenterNet

https://zhuanlan.zhihu.com/p/104573037

论文也撞衫，你更喜欢哪个无锚点CenterNet？

https://baijiahao.baidu.com/s?id=1644905321397514137&wfr=spider&for=pc

CenterNet-FSAF-NMS-single-object

https://github.com/ZY-Russell/CenterNet-FSAF-NMS-single-object

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

https://zhuanlan.zhihu.com/p/66048276

Corner-based对象检测算法三连之——长江后浪MatrixNet

https://www.jianshu.com/p/c94935bab3f8

《FCOS: Fully Convolutional One-Stage Object Detection》阅读笔记

https://blog.csdn.net/watermelon1123/article/details/90045300

DarkNet：

Darknet神经网络框架概述：

https://blog.csdn.net/u010122972/article/details/83541978

DarkNet和YOLO官网：

https://pjreddie.com/darknet/

https://pjreddie.com/darknet/yolo/

YOLOv4源码下载：

https://github.com/AlexeyAB/darknet

YOLO：

YOLO系列算法精讲：从yolov1至yolov4的进阶之路（呕心沥血2万字超全整理，建议收藏！）

https://blog.csdn.net/wjinjie/article/details/107509243

45.2mAP+155FPS! PP-YOLO来了， 精度速度双超YOLOv4

https://blog.csdn.net/Yong_Qi2015/article/details/107625324

YOLOv4

一张图梳理YOLOv4论文

https://blog.csdn.net/DD_PP_JJ/article/details/105744139

YOLOv4 怎么学，看完这一篇就够了！

https://blog.csdn.net/wjinjie/article/details/107445791

重磅更新！YoloV4最新论文与源码!权重！结构！翻译!

https://blog.csdn.net/sinat_39783664/article/details/105728290

重磅更新！YoloV4最新论文！解读yolov4框架

https://blog.csdn.net/Sophia_11/article/details/105726907

一张图梳理YOLOv4论文：

https://www.cnblogs.com/pprp/p/12771430.html

YOLO V4 — 网络结构解析（特详细！）

https://zhuanlan.zhihu.com/p/150127712

YOLO V4 — 损失函数解析（特详细！）

https://zhuanlan.zhihu.com/p/159209199

YOLOv3-Pytorch项目代码解析（详细），项目代码很棒

Github地址: https://github.com/ultralytics/yolov3

【从零开始学习YOLOv3】1. cfg文件解析

https://blog.csdn.net/DD_PP_JJ/article/details/105013056

【从零开始学习YOLOv3】2. YOLOv3中的代码配置和数据集构建

https://blog.csdn.net/DD_PP_JJ/article/details/104019264

【从零开始学习YOLOv3】3.YOLOv3的数据组织和处理

https://blog.csdn.net/DD_PP_JJ/article/details/104709299

【从零开始学习YOLOv3】4. YOLOv3中的参数搜索

https://blog.csdn.net/DD_PP_JJ/article/details/104709330

【从零开始学习YOLOv3】5. 网络模型的构建

https://blog.csdn.net/DD_PP_JJ/article/details/104709403

【从零开始学习YOLOv3】6. YOLOv3中的YOLOLayer解析和推理过程

https://blog.csdn.net/DD_PP_JJ/article/details/104071327

【从零开始学习YOLOv3】7. 教你在YOLOv3模型中添加Attention机制

https://blog.csdn.net/DD_PP_JJ/article/details/104109369

【从零开始学习YOLOv3】8. YOLOv3中Loss部分计算

https://blog.csdn.net/DD_PP_JJ/article/details/105173151

我们是如何改进YOLOv3进行红外小目标检测的？

https://blog.csdn.net/DD_PP_JJ/article/details/108507720

YOLOv3

An experiment of transferring backbone of yolov3 into mobilenetv3

https://github.com/tanluren/mobilenetv3-yolov3

yolov3 yolov4 channel and layer pruning, Knowledge Distillation 层剪枝，通道剪枝，知识蒸馏

https://github.com/tanluren/yolov3-channel-and-layer-pruning

YOLOv3-model-pruning，剪枝

https://gitee.com/dudu00joker/YOLOv3-model-pruning

在 oxford hand 数据集上对 YOLOv3 做模型剪枝（network slimming）

https://github.com/Lam1360/YOLOv3-model-pruning

Model Compression—YOLOv3 with multi lightweight backbones(ShuffleNetV2 HuaWei GhostNet), attention, prune and quantization，剪枝，蒸馏，量化

https://github.com/HaloTrouvaille/YOLO-Multi-Backbones-Attention

YOLOv3模型剪枝，瘦身80%，提速100%，精度基本不变

https://blog.csdn.net/dQCFKyQDXYm3F8rB0/article/details/95041551

（一）使用YOLOv3训练BDD100K数据集之数据集下载

https://www.pianshen.com/article/2479172945/

The backbone is replaced with ShuffleNet v2

https://github.com/ZhuYun97/ShuffleNetv2-YOLOv3

【ONNX】使用yolov3.onnx模型进行目标识别的实验

https://blog.csdn.net/u013597931/article/details/89412272

pytorch-yolov3训练、剪枝、转onnx/tensorrt 加速（记录性文章）

https://blog.csdn.net/hanqu3456/article/details/105492285

YOLOv3剪枝再升级！（后续可参考）

https://mp.weixin.qq.com/s?__biz=MzIwMTE1NjQxMQ==&mid=2247489169&idx=2&sn=f8b1c76a6657b8065d039df704df5913&chksm=96f364c5a184edd3bd3754854c494da4912b6fc90735ee1350ab031862aed5466881f41a68ab&scene=21#wechat_redirect

YOLOv2：

Face Paper: YOLOv2论文详解

https://blog.csdn.net/wfei101/article/details/78944891

【论文学习】YOLO9000: Better,Faster,Stronger（YOLO9000：更好，更快，更强）

https://blog.csdn.net/hysteric314/article/details/53909408

后处理技巧

NMS

一文打尽目标检测NMS——精度提升篇

https://zhuanlan.zhihu.com/p/151914931

一文打尽目标检测NMS——效率提升篇

https://zhuanlan.zhihu.com/p/157900024

IoU

目标检测回归损失函数简介：SmoothL1/IoU/GIoU/DIoU/CIoU Loss

https://zhuanlan.zhihu.com/p/104236411

IoU、GIoU、DIoU、CIoU损失函数的那点事儿

https://zhuanlan.zhihu.com/p/94799295

目标检测之 IoU

https://blog.csdn.net/u014061630/article/details/82818112

目标检测算法之AAAI 2020 DIoU Loss 已开源(YOLOV3涨近3个点)

https://cloud.tencent.com/developer/article/1558533

IoU变形代码实现

darknet with GIoU

https://github.com/generalized-iou/g-darknet

Distance-IoU Loss into YOLO v3

https://github.com/Zzh-tju/DIoU-darknet

语义分割

Detectron2入门教程

https://blog.csdn.net/weixin_36670529/article/details/104021823

One-stage网络：

SSD

MobileNetV3-SSD for object detection and implementation in PyTorch

https://github.com/shaoshengsong/MobileNetV3-SSD

Ssd.pytorch可视化实践

https://zhuanlan.zhihu.com/p/93883352

【目标检测实战】Pytorch—SSD模型训练（VOC数据集）

https://zhuanlan.zhihu.com/p/92154612

目标识别：SSD 论文及pytorch代码学习笔记

https://blog.csdn.net/zxd52csx/article/details/82795104

DSSD(Deconvolutional Single Shot Detector)算法理解

https://blog.csdn.net/zj15939317693/article/details/80599596

RetinaNet

Retinanet目标检测算法(简单,明了,易用,全中文注释,单机多卡训练,视频检测)(based on pytorch,Simple, Clear, Mutil GPU)

https://github.com/yatengLG/Retinanet-Pytorch

Pytorch implementation of RetinaNet object detection.

https://github.com/yhenon/pytorch-retinanet

Two-stage网络：

R-CNN系列算法详解：R-CNN —》Fast R-CNN —》Faster R-CNN 进阶之路

https://blog.csdn.net/wjinjie/article/details/105930512

（RegionProposal Network)RPN网络结构及详解

https://blog.csdn.net/qq_36269513/article/details/80421990

四、全卷积网络FCN详细讲解（超级详细哦）

https://blog.csdn.net/qq_41760767/article/details/97521397

全卷积网络 FCN 详解，用于语义分割

https://blog.csdn.net/tuuzhang/article/details/81004731

阿里巴巴PAI平台

FastNN提供在阿里巴巴集团的机器学习平台PAI

https://github.com/alibaba/FastNN

Alink is the Machine Learning algorithm platform based on Flink

https://github.com/alibaba/Alink

机器学习PAI-人人都用得起的机器学习平台-阿里云

https://www.aliyun.com/product/bigdata/product/learn

刚刚，阿里重磅发布机器学习平台PAI 3.0！

https://blog.csdn.net/cpongo3/article/details/89027125