Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

RGB2Depth #5

Closed
2 of 4 tasks
fzd9752 opened this issue Oct 20, 2017 · 3 comments
Closed
2 of 4 tasks

RGB2Depth #5

fzd9752 opened this issue Oct 20, 2017 · 3 comments

Comments

@fzd9752
Copy link
Contributor

fzd9752 commented Oct 20, 2017

目的:验证现有模型预测深度的可靠性,为是否进一步改良模型提供依据。

  • Quantitative Evaluation, 量化方法,与其他方式对比
    重现已有模型,根据评估误差横向评估
    需要输出图像达到256x256

2.- [ ] Deeper Depth Prediction with Fully Convolutional Residual Networks upsampling to 640x480
https://arxiv.org/pdf/1606.00373.pdf
https://github.com/iro-cp/FCRN-DepthPrediction 代码不完整

4.Learning Depth from Single Monocular Images Using Deep Convolutional Neural Fields https://arxiv.org/pdf/1502.07411.pdf

6.Single-Image Depth Perception in the Wild
https://arxiv.org/pdf/1604.03901.pdf
https://github.com/wfchen-umich/relative_depth
7.Multi-Scale Continuous CRFs as Sequential Deep Networks for Monocular Depth Estimation
https://arxiv.org/pdf/1704.02157.pdf
https://github.com/danxuhk/ContinuousCRF-CNN
选用

@fzd9752
Copy link
Contributor Author

fzd9752 commented Oct 20, 2017

深度图point cloud可视化:

  • 深度 - > 点云

    • 点云转换python代码初步完成
    • 图片点云转换确认
  • matplotlib 可视化

    • 真实3D
    • 我们生成的3D
  • 模型重新训练:

    • 裁剪:图像等比例cropp3张数据图,最大限度保证源环境
    • 预处理训练集:61000张图片
    • 下午开始训练,预处理文件、训练数据、训练配置放在sdc pytorch-cyclegan 下 23_10_train 文件夹内
  • 色温显示深度方法

    • 初步代码确认,展现细节
  • 完善用可视化评测深度预测的方法

    • 多色彩2D展示
    • 真实深度
    • real_2_1.png
    • 生成深度
    • fake_3_1.png
    • 深度数值可视化统计
    • 真实深度
    • real_9_1.png
    • 生成深度
    • fake_23_1.png
    • 3D重建
    • 真实深度
    • real.png
    • 生成深度
    • fake.png
  • 等待新训练模型看效果

VKITTI 深度信息说明

They correspond to the z coordinate of each pixel in camera coordinate space (not the distance to the camera optical center).
camera intrinsic matrix (in pixels, constant, computed from our 1242x375 resolution and 29° fov):

          [[725,    0, 620.5],
 K =    [   0, 725, 187.0],
           [   0,     0,       1]]

点云转换原理 http://www.cnblogs.com/gaoxiang12/p/4652478.html
点云转换python https://codereview.stackexchange.com/questions/79032/generating-a-3d-point-cloud
RVIZ - http://gazebosim.org/tutorials?tut=drcsim_visualization&cat=
@waxz 提供的pyntcloud notebook中呈现3D效果

@waxz
Copy link
Contributor

waxz commented Oct 23, 2017

优化模型的方法
1、深度分割联合预测
http://users.eecs.northwestern.edu/~xsh835/assets/cvpr2015_depth.pdf
2.Single RGB Image Depth and Certainty Estimation via Deep Network and Dropout
3.深度图超分辨
由于深度图一般的分辨率比输入图像小很多,需要采用超分辨率的方法增大分辨率
https://arxiv.org/pdf/1605.09546v1.pdf

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants