CNN阴影去除--DeshadowNet: A Multi-context Embedding Deep Network for Shadow Removal
2017-08-29 13:49
1836 查看
DeshadowNet: A Multi-context Embedding Deep Network for Shadow Removal
CVPR2017
本文使用深度学习CNN网络来进行阴影去除,最大的特色就是全自动的端对端的实现阴影去除。 automatic and end-to-end deep neural network (DeshadowNet)
阴影去除也算是一个老大难问题了,目前存在的方法主要问题有如下三点:
1)Lack of a fully-automatic and end-to-end pipeline
2)Neglect high level semantic information, 目前大部分算法采用 low-level 特征, color ratios,color statistics 。但是阴影同样与 semantic contents 高度相关 (例如 geometry and material)
3)Require specific operation for penumbra regions 对于半阴影区域需要特别操作
针对阴影去除,目前还没有很好的数据库,我们自己建立了一个较大数据库
3 A New Dataset for Shadow Removal – SRD
自己拍照搞了 3088 图像对,主要考虑以下四个因素:Illumination,Scene,Reflectance,Silhouette
我们提出的 DeshadowNet 是 multi-context的,它综合 high-level semantic information, mid-level appearance information and local image details 这些信息来进行最终的预测, 这个 multi-context embedding 是通过三个子网络来实现的: global localization network (G-Net), appearance modeling network (A-Net), and semantic modeling network (S-Net)
G-Net 提取阴影特征表示来描述 场景中的全局结构和 high-level semantic context
G-Net extracts shadow feature representation to describe the global structure and high-level semantic context of the scene
A-Net 提取 G-Net 浅层中的 appearance 信息
A-Net acquire the appearance information from the shallower layer of G-Net
S-Net 提取 G-Net 深层中的 semantic 信息
S-Net acquire the semantic information from the deeper layer of G-Net
本文提出的网络结构
本文提出的网络结构中间结果的显示
网络模型参数设置
损失函数定义
我们采用了 Mean Squared Error (MSE) as the loss function in the log space
Training strategy
为了防止过拟合,我们采用以下训练策略:
1)Multi-stage training strategy 多阶段训练,先分开训练G-Net+A-Net and G-Net+S-Net,然后再整体训练
2) Multi-size training strategy 多尺度训练, coarse scale 64 × 64, medium scale 128 × 128, and fine scale 224 × 224
3)Data synthesis 合成更多的训练数据,60,000 640×480
4)Data augmentation 包括 image translations, flipping and cropping
CVPR2017
本文使用深度学习CNN网络来进行阴影去除,最大的特色就是全自动的端对端的实现阴影去除。 automatic and end-to-end deep neural network (DeshadowNet)
阴影去除也算是一个老大难问题了,目前存在的方法主要问题有如下三点:
1)Lack of a fully-automatic and end-to-end pipeline
2)Neglect high level semantic information, 目前大部分算法采用 low-level 特征, color ratios,color statistics 。但是阴影同样与 semantic contents 高度相关 (例如 geometry and material)
3)Require specific operation for penumbra regions 对于半阴影区域需要特别操作
针对阴影去除,目前还没有很好的数据库,我们自己建立了一个较大数据库
3 A New Dataset for Shadow Removal – SRD
自己拍照搞了 3088 图像对,主要考虑以下四个因素:Illumination,Scene,Reflectance,Silhouette
我们提出的 DeshadowNet 是 multi-context的,它综合 high-level semantic information, mid-level appearance information and local image details 这些信息来进行最终的预测, 这个 multi-context embedding 是通过三个子网络来实现的: global localization network (G-Net), appearance modeling network (A-Net), and semantic modeling network (S-Net)
G-Net 提取阴影特征表示来描述 场景中的全局结构和 high-level semantic context
G-Net extracts shadow feature representation to describe the global structure and high-level semantic context of the scene
A-Net 提取 G-Net 浅层中的 appearance 信息
A-Net acquire the appearance information from the shallower layer of G-Net
S-Net 提取 G-Net 深层中的 semantic 信息
S-Net acquire the semantic information from the deeper layer of G-Net
本文提出的网络结构
本文提出的网络结构中间结果的显示
网络模型参数设置
损失函数定义
我们采用了 Mean Squared Error (MSE) as the loss function in the log space
Training strategy
为了防止过拟合,我们采用以下训练策略:
1)Multi-stage training strategy 多阶段训练,先分开训练G-Net+A-Net and G-Net+S-Net,然后再整体训练
2) Multi-size training strategy 多尺度训练, coarse scale 64 × 64, medium scale 128 × 128, and fine scale 224 × 224
3)Data synthesis 合成更多的训练数据,60,000 640×480
4)Data augmentation 包括 image translations, flipping and cropping
相关文章推荐
- A Multi-task Deep Network for Person Re-identification
- 多尺度R-CNN论文笔记(5): A Unified Multi-scale Deep Convolutional Neural Network for Fast Object Detection
- 人脸关键点: DCNN-Deep Convolutional Network Cascade for Facial Point Detection
- Context-Aware Network Embedding for Relation Modeling
- CANE-Context-Aware Network Embedding for Relation Modeling论文学习
- HD-CNN: HIERARCHICAL DEEP CONVOLUTIONAL NEURAL NETWORK FOR IMAGE CLASSIFICATION(泛读)
- 论文笔记之:Dueling Network Architectures for Deep Reinforcement Learning
- 【论文笔记】Leveraging Datasets with Varying Annotations for Face Alignment via Deep Regression Network
- PS长阴影生成工具 Long_Shadow_Generator_v1.2_for_CS6.zxp
- 多尺度R-CNN论文笔记(1): A MultiPath Network for Object Detection
- Install and Compile MatConvNet: CNNs for MATLAB --- Deep Learning framework
- Deep CNNs for Diabetic Retinopathy Detection笔记
- 人脸关键点:DAN-Deep Alignment Network: A convolutional neural network for robust face alignment
- 用matlab训练数字分类的深度神经网络Training a Deep Neural Network for Digit Classification
- Reading Note: ThiNet: A Filter Level Pruning Method for Deep Neural Network Compression
- Context-Dependent Pre-Trained Deep Neural Networks for Large-Vocabulary Speech Recognition
- 吴恩达 深度学习 1-4 课后作业2 Deep Neural Network for Image Classification: Application
- 目标检测分割--BlitzNet: A Real-Time Deep Network for Scene Understanding
- READING NOTE: ENet: A Deep Neural Network Architecture for Real-Time Semantic Segmentation
- 人群密度估计--CrowdNet: A Deep Convolutional Network for Dense Crowd Counting