Milestone CEO Fights Against CCTV 2.0

"I am going to tell the owner what they can and can not do"

canon "ya fired"

"They [Canon] don’t have access to certain areas of our facilities. They don’t have access to our servers. They will have one seat on the board of directors, ......"

Too funny. Dude they have the keys and that one seat on the board nonsense, that one seat holds the top trump card. Maybe you've been spending too much time in Christiania since Canon purchased your company.

• "An elderly Alzheimer’s sufferer may be prone to repeated, cyclical activities through memory impairment. Getting up to go to the fridge, not realizing why they are there, and repeating the action time and again. Video analytics that can automatically alert a caregiver"

• "Automated analytics software which is designed to detect facial expressions of gamblers to see if they’re happy, or bored, or upset. Imagine sitting at a gaming table, thinking that you’re about to leave, when suddenly a cocktail waitress appears with another of your favorite drinks, and you have a sip, and you think, you know what? I might stick around for another hour."

I really wish people would stop just making up wild claims about what video analytics might be able to do, particularly in cases where they don't appear to be doing any video analytics development of their own.

In many ways his statements above are arguments for the "proprietary jail", though I think that term in borderline inflammatory.

If we're ever going to get to the point that video analytics is capable of detecting these nuanced expressions on a persons face, and making an accurate read of them, and then dispatching a cocktail waitress (really, that's your killer app??) it's going to start off in a situation where you need a tightly controlled integration between the camera, the analytics (wherever they run, doesn't matter) and the VMS.

At VideoIQ, it was *hard* to try and get any of the major VMS's to *want* to do a deep integration. Years later simple things like handling the live meta-data from the analytics still lacked widespread support. I witnessed VMS companies claim all these things were going to be possible, but they didn't appear willing to invest their own efforts to really make it happen. They want the outside vendors to essentially pay money to get developers to add specialized support, or add enhanced features to the VMS API. And of course, every VMS is different, so if you're a smaller startup trying to push the envelop of something like analytics, you probably don't have the resources to essentially duplicate a lot of integration work for each VMS's API.

What does that mean? IMO, it means that we're most likely to see true technological progression on things like "Minority Report" level analytics from companies that can control all the components. They can build the cameras to work with the analytics to work with the VMS, etc. All the API's are in sync, all the teams are working toward a shared goal and making decisions (what DSP should we put in the next-gen camera) ahead of time to converge on a solution that requires all these components working in concert.

Milestone is a great VMS, and I agree that there should be interoperability among devices to give the greatest flexibility. But breakthroughs in things like analytics are not going to come from 3 "independent" companies all somehow converging on a solution separately, breakthroughs are going to come from ONE company recognizing that in order to tackle the problem you need to be able to impact all of the key components.

At some point, when the solution, and the business market are now "obvious", you'll see other companies duplicate what works, or add support for the required API's because they'll need to in order to stay relevant. I was going to post something on another thread, but a close example to this is the fact that Dahua and Hikvision have seemingly come from nowhere in the last couple of years. The industry finally standardized (in a way) on PoE, h.264 streaming and 2-3MP image sensors. There was finally a clear indication of what the market wanted and what could be made in bulk. We got past the JPG vs. MP4 vs. H.264 debates. We figured out what resolution was "enough" for most applications, demand for ONVIF became measurable, and so forth. This was easier to do with cameras because at the core they are simple, passive devices. All you expect of them is a good image at a given framerate.

Analytics is a lot harder. For the examples above there still isn't even what I would consider 100% solid consensus on what kinds of cameras you'd use, or where you'd install them. Would a 360 camera work? Do you need a 4K camera to really see those face details? What about a thermal camera to detect hotspots on a persons face? What about a 360 degree thermal camera? (personally, I think we could wager a rough guess on what would work and what wouldn't, but I'm making a point).

The openness of a platform like Milestone is not going to make any measurable impact on the advancement of the industry into truly new applications. It certainly doesn't hurt, but IMO it doesn't make the list of the top 3 requirements.

If someone developed the analytics that Milestone described above they would have a massive market, huge customer demand, and quite frankly every VMS vendor would be clamoring to add support. You could have the most closed software on the planet, and if the customer said they needed support, that VMS vendor would be forced to add it, because the value of analytics in a case like that would be cause for almost any customer to switch platforms (or, to run two VMS systems).

Maybe Canon will show the Milestone CEO the proverbial "door" for lack of respect of someone who just bought you.

Wireless door contacts, pressure sensors, Fit-Bit, like droid, can tell you a lot about senior care, activities, and the need for help. NFC/Bluetooth can do this now for a few hundred bucks.

Something is amiss here - either Milestone's CEO is confused or Canon has a terrible ownership structure.

I don't think either CEO is confused. I DO believe that Mr. Mitarai has told Lars that he will able to run his company as he sees fit, just like always.

But there is likely an explicit or implicit caveat: "As long as you are making your numbers."

Let's see what happens if and when they blow a couple quarters in a row...

Microsoft (Project Oxford) and other companies (Affectiva) have facial expression algorithms that are being/have been integrated to Gaming applications and market research apps... maybe these use cases are science fiction for Milestone but not for others. I don't believe VMS vendors will take that leap because that solution is targeted at a completely different buyer and industry that these company's don't understand or have the skill set to sell to.

Many companies are jumping on the "Big Data" bandwagon, particularly the PSIM companies; however, even IT companies (who have specialized big data practices) are having challenges monetizing these very involved and protracted sale.

Proprietary jail is a term that has been used long before Milestone coined it (?)

Update: Milestone has announced a case study about a car advertising solution that uses analytics, Axis cameras and Milestone VMS. Basically it detects the model of the car passing and then customizes the advertising displayed.

Video here:

I don't know how well it works or how the costs of operation compare to its return. Even if it's great, the big question for me is why do you need a VMS (any VMS) to do this? The analytic / ad solution can fairly easy connect to cameras directly.