人车密度估计--Towards perspective-free object counting with deep learning

Towards perspective-free object counting with deep learning
ECCV2016
https://github.com/gramuah/ccnn

本文针对人车密度估计问题，主要做了两个工作：1）提出了一个 novel convolutional neural network：Counting CNN (CCNN)，将图像块回归到密度图，2）第二个工作就是提出了一个 scale-aware counting model，Hydra CNN，用于学习 multiscale non-linear regression model

这里我们将人车密度估计问题转为回归问题

3 Deep learning to count objects
3.1 Counting objects model
ground truth density map D 真值密度图由高斯核对人车位置进行卷积得到，有了密度图通过积分得到图像中总的人车数

3.2 The Counting CNN

这个网络使用了两个 max-pooling，输入尺寸是 72x72 ，输出的密度图尺寸是18x18 变为原来的 1/4

Given a test image, we first densely extract image patches
给定一张测试图像，我们从图像中提出很多重叠的图像块，对图像块进行密度估计，再有这些图像块密度图组合为完整图像的密度估计图

3.3 The Hydra CNN
对于一般的基于回归的计数模型，通常需要对输入特征进行 geometric correction， using an annotated perspective map of the scene
为什么需要这个矫正了？主要还是 perspective distortion
Technically, the perspective distortion exhibited by an image, causes that features extracted from the same object but at different scene depths would have
a huge difference in values. As a consequence, erroneous results are expected by models which uses a single regression function

这里我们提出一个多尺度CNN组合学习网络

这个网络结构的设计参考了文献【24】

4 Experiments
Vehicle counting results

UCSD dataset

UCSD dataset

UCF CC 50 dataset

人车密度估计--Towards perspective-free object counting with deep learning相关推荐

[阅读笔记] For：Object Detection with Deep Learning: The Definitive Guide
文章指路:Object Detection with Deep Learning: The Definitive Guide 1 对比其他计算机视觉问题 1.1 图像分类算是计算机视觉里比较著名的问 ...
目标检测：Object Detection with Deep Learning: A Review
文章目录 1. INTRODUCTION 2. A BRIEFOVERVIEW OFDEEPLEARNING 3. 通用目标检测 A. Region Proposal Based Framework: ...
Deep Learning for Generic Object Detection: A Survey -- 目标检测综述总结
最近,中国国防科技大学.芬兰奥卢大学.澳大利亚悉尼大学.中国香港中文大学和加拿大滑铁卢大学等人推出一篇最新目标检测综述,详细阐述了当前目标检测最新成就和关键技术.文章最后总结了未来8个比较有前景的 ...
CVPR2020-Reverse Perspective Network for Perspective-Aware Object Counting：透视感知目标计数的反向透视网络
CVPR2020-Reverse Perspective Network for Perspective-Aware Object Counting:透视感知目标计数的反向透视网络 Reverse P ...
2019论文阅读：SIMCO: SIMILARITY-BASED OBJECT COUNTING
发表于CVPR2019 文章目录文章贡献 1.绪论 2.SIMCO的两个阶段 2.1 SIMCO detection InShape数据集网络结构: 2.2 SIMCO clustering 3. ...
无人机数车--Drone-based Object Counting by Spatially Regularized Regional Proposal Network
Drone-based Object Counting by Spatially Regularized Regional Proposal Network ICCV2017 数据库:https:// ...
人群计数--Cross-scene Crowd Counting via Deep Convolutional Neural Networks
Cross-scene Crowd Counting via Deep Convolutional Neural Networks CVPR2015 本文主要使用深度学习来完成跨场景人群计数 cros ...
【笔记】Comparison of Object Detection and Patch-Based Classification Deep Learning Models on Mid- to La
<Comparison of Object Detection and Patch-Based Classification Deep Learning Models on Mid- to La ...
论文阅读（3）:Image-Based 3D Object Reconstruction:State-of-the-Art and Trends in the Deep Learning Era
论文标题:Image-Based 3D Object Reconstruction:State-of-the-Art and Trends in the Deep Learning Era 论文类型: ...

人车密度估计--Towards perspective-free object counting with deep learning

人车密度估计--Towards perspective-free object counting with deep learning相关推荐

最新文章

热门文章