阅读笔记RGB-‘D’ Saliency Detection With Pseudo Depth
2020-06-25 20:48
190 查看
贡献
1.提出伪深度的测量方法。
2.提出伪深度背景先验。
3.提出基于伪深度的显著性检测算法Pseudo Depth Prior (PDP)。
4.在RGB模型中套用PDP
介质传输模型中的伪深度
介质传输与空间相关,t(x,y):介质传播
图1 介质传播模型
公式1 深度值与传播的对数成反比
对于室内物体,可能出现环境光的干扰,提出“半反相”图像深度(semi-inverse image depth)。
公式2 半反相图像深度
RGB-'D‘显著性检测算法
图2 PDP算法流程图
公式3 融合后的显著性图公式
小于阈值 τ 的显著性值设为0。
实验结果
进一步的实验证明
将方法适配于新提出的RGB显著性模型DRFI和AMC。
相关文章推荐
- 行为识别阅读笔记(paper + parted code):Beyond Frame-level CNN Saliency-Aware 3-D CNN with LSTM for Video Acti
- Frustum PointNets for 3D Object Detection from RGB-D Data 原文 + 阅读笔记
- 论文阅读笔记:Xception: Deep Learning with Depthwise Separable Convolutions
- 阅读笔记4:CSK:Exploiting the Circulant Structure of Tracking-by-detection with Kernels
- 论文阅读笔记:Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
- MaskLab: Instance Segmentation by Refining Object Detection with Semantic and Direction Features阅读笔记
- 论文笔记:Reinforcing LiDAR-Based 3D Object Detection with RGB and 3D Information
- 【论文阅读笔记】CheXNet: Radiologist-Level Pneumonia Detection on Chest X-Rays with Deep Learning
- 阅读图像显著性检测论文四:Saliency Filters Contrast Based Filtering for Salient Region Detection
- 转-Efficient Salient Region Detection with Soft Image Abstraction笔记
- Android官方Training阅读笔记 ---- Building a Dynamic UI with Fragments(Creating a Fragment) (一)
- Perfection and Simplicity A Conversation with Ken Arnold, Part I 阅读笔记
- Tensorflow object detection API 源码阅读笔记:基本类(1)
- 论文阅读笔记:Object Detection Networks on Convolutional Feature Maps
- "Contracts and Interoperability A Conversation with Anders Hejlsberg, Part V" 阅读笔记
- 论文阅读笔记《leaning spatiotemporal features with 3D convolutional network》
- 【论文阅读】Pseudo Mask Augmented Object Detection
- Android官方Training阅读笔记 ---- Building a Dynamic UI with Fragments(Building a Flexible UI) (二)
- 论文阅读笔记:You Only Look Once: Unified, Real-Time Object Detection
- 论文笔记:Faster R-CNN:Towards Real-Time Object Detection with Region Proposal Networks