What is the best demo I can give to Nvidia team to show our skills

Hi Experts

I am planning to buy NVidia TX1 and looking for ideas, which can impress the Nvidia team … I have with me a team who are embedded engineers. As I am not an Algorithm developer, looking for project that enables me to develop tools to test the Nvidia chipset. Also suggest some application where we can showcase working of the chipset drivers. Our skills are Multimedia, Linux Kernel, Device driver development lots of integration and testing skills. Your advice is appreciated.

Hi protocoder, is your aim to have your outfit join NVIDIA’s Jetson Ecosystem for sw multimedia, kernel, and driver services? You could work on the device tree by integrating cameras, SPI/I2C devices and others, and add content to the wiki, that would be helpful. I would suggest looking around the forums to identify areas and devices thay developers are working to integrate into their solutions.

Thank you Dusty_nv.

I agree with you. With our skills, this is probably good.

Can we connect following cameras which we already have to show a PoC…
We have with us one IMX274-MIPI-FMC and rest Rasberry pi Noir cameras… can we do connect all these to Nvidia TX1 and create a project of some integration of cameras …

Will I able to get NVIDIA VRWorks 360 Video SDK stitching algorithms to create some projects.

Thank you in Advance

Hi protocoder, VRWorks isn’t supported/available for Jetson TX1/TX2.

Hi Dust-nv, thank you again for the information.

I thought of this from this link: https://www.ridgerun.com/nvidia-jetson-products-and-services

Please can you pass me to links where I can know what kind of built in algorithms, I can utilize so that I can know work within the capabilities of Jetson TX1 and our skills.
Please advise me

Thank you very much

JetPack is the SDK that supports Jetson and includes the L4T BSP, TensorRT, VisionWorks, the Multimedia API, ect.

Based on your request for 360 stitching, you may be interested to build applications with VisionWorks. It includes GPU-accelerated vision primitives like keypoint extraction + matching that you can use to perform the image alignment.

You may also be interested in our deep learning tutorial for Jetson that you can use to get started making your own apps with TensorRT + DIGITS.

Thank you dusty_nv, I think now I will look into BSP, MM of the TX1 initially and move to visionworks. Thank you for the help. Much appreciated. Thank you admin for marking as answered.

A video walkthrough of natively installing NVIDIA DIGITS on Ubuntu 18.04 LTS is available here:

https://cudaeducation.com/nvidiadigits/

-Cuda Education
cudaeducation.com