Hello trend,
Depth estimation using stereo vision (two views of a scene) is one of the most requested topics on our blog.
Algorithms for depth estimation did not involve machine learning until recently.
But Deep Learning is an idea whose time has come, and its benefits have reached stereo vision. In today's post, we will learn how Deep Learning is used for disparity estimation. We will learn about Pyramid Stereo Matching Network (PSMNet) that is the inspiration for many state of the art networks.
https://learnopencv.com/disparity-estimation-using-deep-learning
and the link to the code is at
https://github.com/spmallick/learnopencv/tree/master/Disparity-Estimation-Using-Deep-Learning
Previous posts in this series are listed below.
- Introduction to Epipolar Geometry and Stereo Vision
- Making a Low-Cost Stereo Camera Using OpenCV
- Depth Estimation Using Stereo Camera and OpenCV
Kickstarter Add-On: $99 OAK-D-Lite (Limited Supply)
As a surprise to our backers, we added limited quantities of OAK-D-Lite at a jaw dropping price of just $99.
Yes, it is the same spatial AI camera that raised more than $1 Million in our Kickstarter campaign last year.
It comes with a 4K RGB camera, a stereo pair for depth estimation, a neural inference engine powered by Intel's Myriad X capable of 4 Trillion Operations Per Second.
OAK-D-Lite requires no prior knowledge of AI or spatial AI. Its Python API, documentation, and tutorials will leapfrog you effortlessly into the fascinating world of spatial AI.
Learn more about OAK-D-Lite here
How to get the OAK-D-Lite Add-on?
If you have already backed our campaign and want to change their pledge to add this add-on, please check out this video for instructions.
If you want to back our campaign and also add this add-on, please check out this video for instructions. In the same video we also show how to get the add-on without backing the course by pledging $1.
The price will increase to $109, before settling at $119 during this campaign.
| Kickstarter : DL with Tensorflow & Keras |
Kickstarter Stretch Goals
323 backers who helped us raise more than $131,210 in our Kickstarter campaign for Deep Learning with Tensorflow & Keras.
Thank you so much!
We have announced new stretch goals that add even more emphasis on model deployment, and building applications.
- Deploying models as a web app using Streamlit.
- Deploying models as a REST API using FastAPI.
- Introduction to Google Cloud API for computer vision.
- Building Computer Vision applications using MediaPipe.
| Kickstarter : DL with Tensorflow & Keras |
Cheers!
Satya
No comments:
Post a Comment
Note: Only a member of this blog may post a comment.