Ex Googler Startup, Camio, Targets SurveillanceBy: John Honovich, Published on Dec 04, 2013
A team of former Google engineers are targeting the surveillance market with a startup focusing heavily on analytics and search. In this note, we examine Camiolog's product positioning, comparing it traditional surveillance and rival new entrants like Dropcam.
Camiolog is a service that processes and records video in the cloud. The main differentiators are (1) performing video analytics automatically on all incoming video and (2) a search optimized user interface.
Roughly speaking, Camiolog is like 3VR, Briefcam and Siri combined. Camiolog uses thumbnails, generated and sorted based on advanced motion detection, color recognition and zones. It creates synopses, which they call 'camios' that are video/image series summarizing the key activities over a period of time. Finally, in their mobile app, voice commands can be used to find relevant video (e.g., "Show me Tuesday front red interesting motion events").
The image below shows a series of thumbnails filtered to specific events:
Unlike almost all video systems, the focus here is on search with minimal options for live monitoring.
Instead, they add features like zones, where motion detection areas can be labelled, like so:
These labels (driveway, walkway, lawn) are what enable users to do google / siri like searches. The system does not know what a 'lawn' is but it has a record that maps your entry for 'lawn' to a specific zone on a specific camera (compare to ipConfigure's similar approach).
Additionally, the system automatically catalogs colors. Here's two of their marketing examples:
Of course, in these two, the white and blue trucks are much different colors that the surrounding scene. Something green or grey would have more difficulty with matching accurately.
Here's a short video clip that shows the whole search process in action, from typing in a free form query to reviewing results:
Additionally, the same process can be used to set up alerts / notifications for events that match specific queries / parameters.
Camiolog supports plug and play connectivity from Sharx [link no longer available] (1 indoor model at $199, 1 outdoor model at $349). 3rd party cameras are supported though one needs to manually set up integration via FTP or HTTP (see multi-step Axis process [link no longer available]).
The service is targeted at residential / consumer users and costs $16.50 per month. There is no per camera fee but it is restricted to cameras at a single site and only motion triggered video of certain sizes.
Additionally, a free Web cam option is useful for trying out the system and can also be used for basic monitoring. Also, a free plan allows for live viewing and 8 snapshots per hour from a web cam, but without the analytics, it does not have much use.
Camiolog sees its core competitive advantage in search - the ability to quickly find meaningful results in lots of videos. In this video, they make the case of why their approach is far better than traditional surveillance as well as Dropcam:
The main problem we see is that search is not a core need for most surveillance users. Even for commercial / enterprise users, this is not the case. I found this personally when I was at 3VR and the ongoing struggles of companies like 3VR and Briefcam to become major players further validates this challenge. And with consumers, it is even harder to imagine that search will be such a driving factor.
By contrast, Dropcam's adoption is clearly driven by its ease of setup and use (as we have shown in our test reports). This has always been a huge challenge for consumers and SMB who overwhelmingly have no idea how to setup FTP, HTTP, port forwarding, etc. We expect the consumer surveillance market to continue to be driven by providers who make it easy to buy and deploy their products. Very few will choose a more complicated buying / setup process for greater video intelligence.
At the same time, what Camiolog is doing on a technical level is pretty amazing. To deliver advanced VMD, color analytics, synopses, natural language search and voice activated commands for less than $20 per month, no server required, is groundbreaking. However, positioned at the residential market as is, it faces significant challenges in achieving product / market fit.
All that noted, it is promising to see a team of engineers developing real technology that may eventually evolve / mature into something more productive to regular users.