本文结构:
0 前言
1 论文经常提及的几个知识点
2 最新论文
3 one-stage detectors
4 two-stage detectors
5 其他
0.前言
目标检测算法盘点
机器之心:从RCNN到SSD,这应该是最全的一份目标检测算法盘点zhuanlan.zhihu.com先上三张图直观感受一下目标检测算法在准确度(AP)和速度(speed)的对比
1.论文经常提及的几个知识点
- 评价指标AP(mAP)
- 边框回归(Bounding Box Regression)
- 非极大值抑制(non maximum suppression)
- OHEM(Training Region-based Object Detectors with Online Hard Example Mining)
论文地址:https://arxiv.org/pdf/1604.03540.pdf
论文解读:
论文笔记 OHEM: Training Region-based Object Detectors with Online Hard Example Miningblog.csdn.net- RPN(Region Proposal Network)
2.最新论文
CVPR2018 目标检测(object detection)算法总览blog.csdn.net3.one-stage detectors
- SSD(Single Shot MultiBox Detector)
论文地址:https://arxiv.org/pdf/1512.02325.pdf
论文解读:
我是小将:目标检测|SSD原理与实现zhuanlan.zhihu.comSSD: Single Shot MultiBox Detector翻译(包括正式版和预印版)(对原文作部分理解性修改)
方良骥:SSD 系列论文总结zhuanlan.zhihu.com- YOLO(You only look once)
论文地址:
You Only Look Once: Unified, Real-Time Object Detection
YOLO9000: Better, Faster, Stronger
YOLOv3: An Incremental Improvement
代码地址:
keras版本:qqwweee/keras-yolo3
论文解读:
YOLO(You Only Look Once)算法详解blog.csdn.netkeras源码解读:
探索 YOLO v3 源码 - 第1篇 训练juejin.im 探索 YOLO v3 实现细节 - 第2篇 模型juejin.im 探索 YOLO v3 实现细节 - 第3篇 网络juejin.im 探索 YOLO v3 实现细节 - 第4篇 数据和y_truejuejin.im 探索 YOLO v3 实现细节 - 第5篇 Lossjuejin.im 探索 YOLO v3 实现细节 - 第6篇 预测 (完结)juejin.im- RetinaNet(Focal Loss for Dense Object Detection)
论文地址:https://arxiv.org/pdf/1708.02002.pdf
论文解读:
论文阅读: RetinaNet - Online Notes - CSDN博客blog.csdn.net- CornerNet
- CenterNet
4.two-stage detectors
- R-CNN(Region-based Convolutional Neural Networks)
论文地址:rbgirshick/rcnn
- Fast R-CNN
论文地址:https://arxiv.org/pdf/1504.08083.pdf
- Faster R-CNN(Towards Real-Time Object Detection with Region Proposal Networks)
论文地址:https://arxiv.org/pdf/1506.01497.pdf
论文解读:
Object Detection and Classification using R-CNNswww.telesens.co- Mask-RCNN
论文地址:https://arxiv.org/pdf/1703.06870.pdf
论文解读:
Mask-RCNN技术解析 - 跟随技术的脚步-linolzhang的专栏 - CSDN博客blog.csdn.net- Cascade RCNN
- TridentNet
5.其他
- FPN(Feature Pyramid Networks for Object Detection)
论文地址:https://arxiv.org/pdf/1612.03144.pdf
论文解读:
FPN(feature pyramid networks)算法讲解blog.csdn.net- FCN(Fully Convolutional Networks for Semantic Segmentation)
论文地址:https://arxiv.org/pdf/1411.4038.pdf
论文解读:
全卷积网络 FCN 详解 - 代码学习者coding - 博客园www.cnblogs.com