计算机视觉论文笔记

写在前面

我是使用notion来记录我的学习笔记，接下来我会陆续把自己学习过程中看到的知识点、论文笔记等发到博客中，如果有哪里出错，欢迎大家指出（实时更新）！

论文笔记：

一、图像分类

1. ZFNet（Visualizing and Understanding Convolutional Networks）
2. GoogLeNet（Inception v1）
3. Inception v2/v3
4. ResNet（Deep Residual Learning for Image Recognition）
5. Inception V4 && Inception ResNet
6. MobileNet V1（MobileNets：Efficient convolutional neural networks for mobile vision applications）
7. Xception（Xception: Deep Learning With Depthwise Separable Convolutions）
8. MobileNet V2（MobileNetV2: Inverted Residuals and Linear Bottlenecks）
9. MnasNet（MnasNet: Platform-Aware Neural Architecture Search for Mobile）
10. MobileNet V3（Searching for MobileNetV3）
11. ShuffleNet V1（ShuffleNet: An Extremely Efficient Convolutional Neural Network for Mobile Devices）
12. ShuffleNet V2（ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design）

二、目标检测

2.1自然场景中的目标检测

1. YOLO v1（You Only Look Once: Unified, Real-Time Object Detection）
2. YOLO v2（YOLO9000: Better, Faster, Stronger）
3. YOLO v3（YOLOv3: An Incremental Improvement）
4. YOLO v4
5. YOLOX（YOLOX: Exceeding YOLO Series in 2021）
6. RCNN（Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation）
7. SPP-net（Spatial Pyramid Pooling in Deep Convolutional Networks for Visual Recognition）
8. Fast R-CNN（Fast R-CNN）
9. Faster R-CNN（Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks）
10. FPN（Feature Pyramid Networks for Object Detection）
11. RetinaNet（Focal Loss for Dense Object Detection）
12. CornerNet: Detecting Objects as Paired Keypoints
13. CenterNet：Objects as points
14. SSD: Single Shot MultiBox Detector

2.2 遥感图像目标检测

1. R2CNN-NOTA（R2CNN++: Multi-Dimensional Attention Based Rotation Invariant Detector with Robust Anchor Strategy）
2. RoI Transformer（Learning RoI Transformer for Oriented Object Detection in Aerial Images）
3. RRPN：Rotated_text_dectection（Arbitrary-Oriented Scene Text Detection via Rotation Proposals）
4. DOTA（DOTA: A Large-scale Dataset for Object DeTection in Aerial Images - Dataset）
5. R³Det（R³Det：Refined Single-Stage Detector with Feature Refifinement for Rotating Object）
6. S²A-Net（Align Deep Features for Oriented Object Detectio）
7. DMNet：Density Map Guided Object Detection in Aerial Images
8. DAL：Dynamic Anchor Learning for Arbitrary-Oriented Object Detection
9. ClusDet：Clustered Object Detection in Aerial Images
10. BBAVector：Oriented Object Detection in Aerial Images with Box Boundary-Aware Vectors
11. ReDet: A Rotation-equivariant Detector for Aerial Object Detection

2.3 小物体目标检测

• 小目标（small）：$area<32^2$
• 中目标（medium）：$32^2<area<96^2$
• 大目标（large）：$96^2<area$

1. QueryDet：Cascaded Sparse Query for Accelerating High-Resolution Small Object Detection

三、图像分割

3.1 传统方法

1. 基于边缘的图像分割方法
2. 基于阈值的图像分割
3. 基于区域的图像分割

3.2 语义分割

1. FCN（Fully Convolutional Networks for Semantic Segmentation）
2. U-Net: Convolutional Networks for Biomedical Image Segmentation

3.3 实例分割

暂无

四、目标跟踪

4.1 单摄像头单目标跟踪

1. 视频单目标跟踪研究进展综述

4.2 单摄像头多目标跟踪

1. FairMot:On the Fairness of Detection and Re-Identification in Multiple Object Tracking
2. TraDeS:Track to Detect and Segment: An Online Multi-Object Tracker
3. QDTrack:Quasi-Dense Similarity Learning for Multiple Object Tracking
4. CenterTrack：Tracking Objects as Points
5. Multiple Object Tracking: A Literature Review
6. Persistent Tracking for Wide Area Aerial Surveillance
7. MeMot : Multi-Object Tracking with Memory

4.3 多摄像头多目标跟踪

1. Branch-and-price global optimization for multi-view multi-target tracking
2. Hypergraphs for Joint Multi-View Reconstruction and Multi-Object Tracking
3. Multi-Camera Multi-Target Tracking with Space-Time-View Hyper-graph
4. Vehicle Re-Identification with the Space-Time Prior
5. Features for Multi-Target Multi-Camera Tracking and Re-Identification
6. Real-Time Multi-Target Multi-Camera Tracking with Spatial-Temporal Information
7. CityFlow: A City-Scale Benchmark for Multi-Target Multi-Camera Vehicle Tracking and Re-Identification
8. People tracking in multi-camera systems: a review
9. Cross-View Tracking for Multi-Human 3D Pose Estimation at over 100 FPS
10. The MTA Dataset for Multi-Target Multi-Camera Pedestrian Tracking by Weighted Distance Aggregation
11. Pose-Assisted Multi-Camera Collaboration System
12. ELECTRICITY: An Efficient Multi-camera Vehicle Tracking System for Intelligent City
13. Multi-Camera Tracking of Vehicles based on Deep Features Re-ID and Trajectory-Based Camera Link Models
14. TRACTA
15. City-Scale Multi-Camera Vehicle Tracking by Semantic Attribute Parsing and Cross-Camera Tracklet Matching
16. Real-time 3D Deep Multi-Camera Tracking

4.4 Re-Identification

1. In Defense of the Triplet Loss for Person Re-Identification

五、计数

5.1 基于密度估计的计数方法

1. CSRNet: Dilated Convolutional Neural Networks for Understanding the Highly Congested Scenes
2. 基于深度卷积神经网络的田间麦穗密度估计及计数

5.2 基于目标检测的计数方法

1. Vehicle counting and traffic flow parameter estimation for dense traffic scenes
2. 基于改进实例分割算法的智能猪只盘点系统设计
3. A machine vision system to detect and count laying hens