Test please !
$38+ Million Funding Powers VMS Challenger IronYun
VMS and video analytics have received little funding this decade.
However, one Taiwan startup, IronYun, has bucked this trend, with a relatively massive amount of funding for an offering that combines deep learning analytics, multi-tier storage, private cloud, and big data. They call it a "Video Operating System" and say it has VMS functionality, real-time analytics and a Google-like video search interface for a true "complete" platform.
In this note, we review the offering based on a discussion with IronYun CEO Paul Sun [link no longer available], analyzing its potential in the market vs established VMSes and analytics.
IronYun **********
*** ******* *** ********** ************ ** May ****, *** ******** *********** ***** 4 ***** *** ******, ***** ** ******* **** ******'* MIT. *** *** *** * ***** background ** ********** ********, *** *** also *** ******** ** *** ***** computing ****** ** ****. ********* ***** are ***** * ***** *********, ****** in ******, *** ******* ** *&*. IronYun ** * ******* ** ***'* efforts ** ****, ***** ** *** they ******* ** ***** * ******* product ** ****** ****** * **** of ***** ******** *************.
***** ******** *** ** ***** ***, the ******* *** ****** ******* $** to ***** *** ******* ** ****** globally.
CityEyes ******* ********
********** *** **** ** *******'* *******, which ************ ******* ********* **** ***** typically **** ** ********* ******* ** a *********** ******. ***/*** *********, *******/***, and ********* *** ********* ** ** integrated *******. ************, ***** ** * deep ******** *********, ***** ******* ****** make *** ****** **** ** * video ****** **** **** * ****** VMS/recorder ******. *** ****** ** ***** to ** **** ** * **** platform *** ** *** ****** ******** to ** **** **** *** *** analytics ********* ***** *******.
** ** ********* ** ** *********, pre-loaded ******** ** * ****-***** ******.
Deep ******** *********
**** ** *******'* *&* (**** **** research, *** *** ******* ****** ***** incorporation) *** **** ******** ** ******** and ******** *** ******, **** ***** own ****, ** ** **** ** automatically ********* *** *** ******* ******* with ****-****. *** ****** ********* *** over *** ****** ******* **** ** can ********* ** * *****, ***** includes *** ******* **** "******" ** "car", ****-****** ***-*******, **** "***", "*******" and "***" *** ********, *** ****-****** ones **** "****** *****", "*****" *** "TV/monitor". ******* **** *** ****** *** recognize ***** ******* **** ******** ******, and **** ******* ***********, **** ** they *** **********. **** ** * significant ******* **** ********* ******* **** rely ** ****** ** ********* *** objects ** ******* *** **************.
Video ******/******
******** *** * ******-**** ****** *********, where ***** *** ****** *** ****** that ******* * **** ***********. *** following ******* ***** ********* *** ***** with ** ***** * ***, *********, and ****** ** *** *****.
***** ********, **** ** ****** *****, can ** ************ ** ****** ******* with * **** *** *** * people *******, ** ** *******.
***** *** ** ***** ***** *** same ***** ********* ** *** ****** interface, ********* ** *******. **** ****** for ****** ***** **** * ****** crossing * ********, ** **** ******* rules **** * ****** ******* ********* the ***** *********.
** ******** ** ****** ************** *********, there ** **** ** **** ******, allowing *** ****** ** **** ******* plates *** ***** **** ** * database, ******** ***** ** **** ** vehicles ***** ** ***** ***** ****** being **** ****** ******** *******.
'Private *****' *** ****** *****
***** ******* ******* **** ** * 'private *****', ** ** *** ***** deployed *** **** ** *****. ******, it ** **** ******* ** ********* to *********** **** **** *****-** *********.
************/*****
******* **** **** *** ************ *** setup ** ***** ********* *** ******* to * *** *********. ******* *** added ** *** ****** *** *** deep-learning ***** ******* ******* *************. ***** purchase *** ********* *** ******* **** IronYun ** ***** ** ** **** to ********* **** ***** ******* *******.
*******
******* ** ***** ** ******** *** total *******, * *** ****** ****** with ~**** ** ******* ***** ** ~$25,000 ********* ** *******. ********** ****** licenses *** $***/*******. *** ****** *** MSRP, *** ******* ***** *** ********** pricing.
Lowest *** ******
** *********** *** **** ****** ********** of ******* ***-**** *********** **** * single-source ********, ******* **** **** *** deliver *** ****** *** **** ******** to *********** *********.
Target ********
******** ** ******** ** ***-**** ***********, typically **** ***+ ******* *** * full-time ******** *****. *******, **************, ******** malls, *** ******* ********* *** ***** IronYun **** **** **** **** *** most *****, ** **** ***** **** a **** ** ****** ******* ***** amounts ** ***** ** **** ******** events.
Existing *********
*** ******* *** ********* ** ** countries, ********* ** ***** ******* *************, though **** *** ** **** (**************). Customers ******* ***** ************* **** *******, airports, ************, *** ******* ************* **** typically **** ****** ****** ****** *** large ******* ** ****** ********.
Sales *******
******* ** ********* ** ****, ***** they **** ******** *********** *** ***********. They *** ********* **** ***** ******* and, ***** **** **** * ****-**** salespeople, *** ***** ******* ***** **** say **** *** ********** ******.
Competitive ********** ** ********** ***** / *********
** **** **** ******** ** * VMS ********* **** ******** ********* *********, making ** ********** ** * ***** 500A **** *** ***** ********* (*********** a *********** ** ******** *** *******), or ** ** ******** *** **** Avigilon *******, **** ******** ***** *************** ******.
*******'* ******* ************** ** ***** **** learning ********* *** ******* ** *********** classify * **** ******* ** ******* in * *****. ********* **** ********* with *** ******* ***** ** ************ may **** ******* **** ********** **** established *****.
Outlook *** *******
***** ********* ***** **** ********** *******, as **** ******** ***** ** ****** for **** ** ****** **************** ****** and ******** *******. *******, ***** ********* are ***** ******** ** ** **** a ****-****** ******** **** * ***** company, ******* **** **** ***** **** operation ** ***** ****** ****** *** easily ** *** ******* ** *** successful ** *** **** ***. **** may ******* * ********* *** ******* with ***** ******* ********** ***-**-*** *******, requiring ***** ** **** * ***** leap ** ***** ** ******* *** majority ** ***** ***** ************** ** the ***** ** * *******.
This is a low priority to test, as we have little evidence to show how well it works and since this post has gotten minimal read levels, indicating low comparative interest.
I have a test unit inbound. I am primarily interested in testing deep learning. Will report if I find it useful.
deep learning with shallow cyber defenses does not an interesting solution make. make sure it's safe before you fall in love with the AI whispering alarms in your ear.
Would be quite interesting for the Middle East market where Asian products generally do quite well.
The name of this company rang a bell to me. Just checked Linked in emails and found this:
"I am the President/CEO of a cloud computing and big data video surveillance startup. I would like to speak to you about a ground floor, exciting high growth opportunity at our company. Please bounce me an email with your contact info.
I look forward to speaking with you.
Best,
Paul Sun
President/CEO
Ironyun
email: paulsun@IronYun.com"
Note: I work for an Analytics company, though I would rather not say which one as it has little relevance to this topic.
You email must be faster than mine. I just got email from Mr. Sun the other day.
I just got invited to be a Senior Board Adviser - anyone else?
Paul Sun and Tom Ma are gentlemen, I wish them luck!
Paul Sun gave me a rundown of their product at ASIS 2016. I have to admit I was impressed and see this type of deep learning analytic as the future for security. I bring it up to anyone in earshot whenever the topic of video analytics comes up. I look forward to seeing where this solution goes in North America
A fully integrated system has the benefit of being homogeneous, with lest complexity in setup probably and simplified support. (One call support). The benefit of a modular design of course being you have greater choice, from best in class to cheapest in class, depending on needs, and not all your eggs are in one basket.
Cameras are added to the system and the deep-learning video tagging happens automatically.
That means CPU power is immediately being siphoned off for the analytics. I wonder what affect that has on number of supported cameras or stream throughput.
It doesn't sound like anything another already established VMS company couldn't also do. Maybe if it starts making inroads, it can kick up more innovation in the industry overall to try and stay ahead or keep up.
Does anyone know how this system compares to other Video Analytics like Avigilon Rialto, Agent VI, ect...?
It is more search oriented than alarm oriented. It would be a closer comparison to something like Camio than Rialto.
So my understanding of this system is that it's primary purpose is to create metadata on the video footage fed into the channel and store it in a more searchable manner.
Would it still be fair to make the comparisons between IronYun and Agent VI since that is similar to what Agent VI does without the "deep learning" portion of the algorithm?
If you are comparing to the savVi Video Search then yes, though IronYun does not offer the event search parameter.
Okay. It will be interesting to see were the go then. Thanks for the information.