Actions
Forum Thread Details
  • Replies 7 replies
  • Subscribers 83 subscribers
  • Views 2187 views
  • Users 0 members are here
Related

Nvidia Jetson Nano: AI and Deep Learning

ipv1

I am working on Machine Vision Applications and OpenCV is my current tool. I was not considering deep learning till I saw the Nvidia Jetson Nano Board which is good on paper.

 

I am awaiting some funds from a few clients in India and will be getting the board myself but before I do, I wanted to know if someone has any experience with TensorFlow, OpenCV or the like with Nvidia boards?

 

Thanks

  • in reply to Jan Cumps

    Hi Jan, I like when you joke image

    As you know (or maybe that the liter of beer past summer in Bruxelles forgot you our nice meeting) it is ten months I develop as a job almost exclusively with OpenCV and custom developed algorithms...

     

    I am sorry that until yesterday I was almost far from the community, due to the Picasso Challenge and the Art-a-Tronic exhibition (when I should expect you to come here??) but reading this post of Inderpreet I have some perplexities on the Depp Learning approach as the best possible in AI. Concerns that grown during these last months just working on a different approach. Unfortunately, instead, I have no experience of the Nvidia new boards, that sounds very promising for AI and image processing but not only.

     

    As some algorithms, I have developed are proprietary in the specific context of the research I am working on, most of the knowledge I acquired and the theories and custom approach I have used I have no difficult to share and discuss if these can be helpful in some way.

     

    Enrico

  • in reply to balearicdynamics

      wrote:

     

    Hi Jan, I like when you joke image

    As you know (or maybe that the liter of beer past summer in Bruxelles forgot you our nice meeting) it is ten months I develop as a job almost exclusively with OpenCV and custom developed algorithms...

     

    ...

     

    Enrico

    I thought I remembered that you abandoned OpenCV for the project because it wasn't up to the job.

  • in reply to Jan Cumps

    True Jan, but then investigating I saw that the real wrong approach was using the OpenCV neural network / Deep learning features. That is almost poor in the OpenCV 3 library. Not to blame them but this feature is just a small toy respect the specialized ones like the open source platform mentioned by Inderpreet. The point is using OpenCV only, as of a good low-level image processing. Then - IMHO - I think it should not be considered as deus-ex-machina.

     

    Enrico

  • in reply to balearicdynamics

    My problem statement involves monitoring objects in varying light conditions as well as very distant objects from the camera. A number of standard information like, identification of object, motion of object and location on the image are required.

     

    a. Any and all help would be greatly appreciated. Any knowledge, theory and/or custom approaches and test case information would be very useful and I would be highly obliged.

     

    b. Can you explain what you mean by OpenCV only  as a good low-level image processing use?

     

    Thanks image

  • Hi , you might get better responses to this if we move it to the Artificial Intelligence area?

  • in reply to cstanton

    I am unable to start a discussion in the space. Can you move the current topic or help start a new one?

    Thanks