NVIDIA For Video Surveillance / Physical Security Profile
NVIDIA has become one of the biggest hardware providers for AI applications but how is NVIDIA positioned for video surveillance and physical security?
While NVIDIA is most well known for its large-scale GPUs, video surveillance is increasingly adding video analytics to cameras. Where does NVIDIA see its role within the physical security market?
IPVM interviewed NVIDIA VP and GM of Embedded and Edge Computing, Deepu Talla, Ph.D., and NVIDIA PR of Autonomous Machines, David Pinto, and discussed NVIDIA's current focus and future outlook concerning the physical security industry including surveillance cameras and robots.
Executive *******
****** ******* ** ********** ******** *** software ************ *** ******* ********, ********* security *** ************, *** ********** **** they ** *** ***** *** **** to ********** ** ***** **-***** ******* (as *.*., ********* **** *********, ******, and ******** **).
Competitive ********
****** ********* *** ***** ** ********* computing ********* **** *****, ******** *********, and ******** *********. ***** ********* ***** on *********** ********* *** ********, ****-*********** computing, *** **, ******** *** ******* purposes *** ******** **********.
****** ********* ******** ******, ***** *********, and ****** *********** ***********, ****** *** been ********** ******** ******** ****** **********. **** ******** ******** * ***** for ** ****-**-***** *********, ** ****** believes **** ********* *** * *** of ***********, ********* ******** ***** ***** towards ****-*********** ********* (***).
* ***** **'* ********** ******* ***** the ***** ** *** **** ** very *********** *** **** **** **** exactly ****. ** **** **** **** focused ** ***** **** *** ****, servers, *** **********.
***** **** **** ***** ***** *** edge ********* ** ** *** ****** product **** *** ***** ********** *** products *** ******* *** **********. ***** emphasized **** ****** **** *** **** on ********** **** ********* *** * single ******.
** **** *** ******* ** ****** camera ** ******* *** ********. ** you **** ** *** ******* ********* in *** ** *** **** ** edge *********, ******* **'* *** ****** product **** ** ******* **'* *** other ********** *** ******* **** ** appliances *** *******, ** ******** *******, like ******* ** ******** ** ****** or ********** *************, *** ** **, so *****.
***** ******* ********* **** *** ***** of ****** *** ******** ** *** only ** ********* ******** *** **** by ********* ******** *********.
*** ***** **** ** ******* ** far ****** *** ******** ********, **** of *** ********** ******** ***** **** our ***********, ***** ********** ******** **** runtime ************ **********, **** ******** ** Maxine. *** **** ** ***** **** and **** *** ******** *********** ***** and **** *** ****** ** ****** for ********* ** ****** ***** ** models, *** *** **** *********** ** the ******.
Metropolis ********
************ ******'* *********** *********, **** ******* developers *** ********** ** ****** ********* involving ****** **** *** **. ********** platform ********** **** *******, *** *******, and ***** *******. ***** ******** ********** use *****:
**'** ******** **** *********** ********* **********, for ***** ********* **** **** **** multiple ********** *** ******** ***** *** that **** ** *** ********** **** use *** *********, ******* **'* *** underlying ******** *** **** *********** ******** tools **** ** ******* *** ***** applications. ** *** * *** ** deployments ** *******, *** *******, ** course, ** **** *** ** **** the ***** *********. ** *** * lot ** ******** ** ****** *** self ********, ********** ********, *******, *** all ***** ***** ** ******. ** see **** * *** ** ********** inspection **** ** ************ **** ******* manufacturing *******. *** ** ******, ********, which ** *** ***** ********.
***** ********* **** ***** ** ** increasing ***** **** **** ** *****, especially *** ******** *********, ******** ******** with ******** ******* ** ******, *** anonymity ** ****** *********.
**** ************* ********** ****** *******************,*******,***** ********,*********,********,*****,******* ***,******, ************.
Jetson **** *** ****
****** ****** *** ****** ******** ** *** edge ********* ********* ** ******, *********** six *** ********. ****** **** **** is *** ******** ****** **** ******, that *** ******* ** ** ** TOPS (****/******** ********** *** ******) ** AI ***********, *** ****** ** $***. Jetson *** **** ****** *** **** of *********** ** * $*** ***** point, ***** ***** ** *** ***-********** edge ****** ** *** ****** ******. These *** **** ********* ***** ********* computing ***** *** ********* ********* ** devices **** ******* ****** ** *********,*** ********, *** ******* *******.
Competitors *** **-***** *********
** *** **** *** *****, ***** analytics ********* ****** ***** ************ ******* has ********* *************, ****** ******* ***** NVIDIA, ** ** ******* ** *** most ******** ****** ********* ********. ******* ************* *** ***** ***** own **** ******** ***** (*.*., **** ARTPEC-8 ** ****** *******) ** ********* from ****** (*.*., **** ******** *********, used ** ********, *****, *-***, *** a **** **** ** ***** ****** providers), **** ************ ** ******* *** have "****" **** ******** *********.
****** ******* ** ** **** ******* in ****** ** ***** ***** ******** applications, ********** **** ********* **** ******* analytics ****** *********** ****** *** ******* classification ** **** ** ********** **-************** across ******** *******.
NVIDIA's ************ *** ******
**** ****** ***** *** **** ******** companies ********* ********* *** ******** ********, Talla ********* **** *** ** **** one ******** ******** ********* ****** ***** with, ******** ****** **** ********* * "minuscule" ********** ******** ** *** ******. Talla ********* **** ********** ***** ************* face:
** ***** ** ******** ******, *****'* a *** ** ********** **** *** still ** ** ****** *** **** are ******* ****** *** ******. *** of *** ******* ********** **** ** had **** ******** ******** *** ****, you ****, ** ******** ***** * long **** ** ***** ***** ****** and ******* **** **** *** ********* them, ********** **** ** ******** ************, and ********* *** **** **** *** amount ** ******* *** **** ** actually ** *** **** ******** ** still * *********.
** ********** ***** ********** ** "**'** got ** *********** *********, ** *** a ********* *********. **'** *** * time ********* ** ***** ** *********** and *** **** ** *** ****** and ******* ** *** ** ****."
***** *** ****, ** ********* *** * ******** that **** *******/*********** **** ******* ** analytics ************. **** ** **** **** talking ***** ***** *** *** *** not **** *** **** ****** ** AI (**, **** ********, ****** ****, etc).
** ** *******, ** *** ************ and ********** ** ***** ********** ********, there **** ** * ****** ****** (especially ***-**** *** *** ****-*** ** existing *******), *** *** ***** **** shift.
****** ******** ******** ********** *** *.*, ** ***** **** *********
- *****-****** ********
- **** ****
- ****** ******* ****** ****** *******
- ***** ****** *******
** *** ********** ****, ****** ********** ****-**-***** ***********.
**** **** **** ******* ********** *** analytics ********** **** **********
** *** ***** ***** ******* ** NVIDIA ********* **** **** ******** *** adoption ** ***** ******* *** ********?
* ******* ** *** ****** ***** GPU's ** *** *******--****'** ******, ********** better **** *** **** ****** ***'* of ******** ***'* *** *** **/*** partners * **** ** *** ************ in **** ******* ** ***'* * revenue ********** ****** ** *** *******.