Potential applications for Digital Video Analysis and a decentralised business strategy to develop it

Digitization of our society is apparently an unstoppable and speeding trend, From Hotmail, Google library, to recreating an entire university inside a computer by Microsoft’s’ WorldWide Telescope. Digitization are driven by different technology “engines”. I am using the terminology “engines” because some technology has so much potential for spin-off applications that it should be view as an engine of digitization. Personal computer was the first major engine that let us digitize our information. Internet was the second major engine that revolutionized information distribution. I believe that next engine of digitization is (for the lack of knowledge of a better term) Digital Video Analysis (DVA) and image analysis, which are also the subject and inspiration of this post. In this post i will attempt to present these following thoughts:

1. What are some of the potential application of DVA
2. What kind of innovation strategy is best able to generate vast amount of product ideas.
3. What kind of business strategy is best able to commercialize these ideas.


By DVA, what I am referring to is the extraction of meaningful information from videos, similar to image analysis but for video. Second to words, visual images are our largest reservoir of information, it has been largely untapped because unlike words it is very difficult to capture, sent, store and analysis large quantity of it in sufficient quality. But that’s changing with the advent of high resolution video cameras, optical Internet, massive storage and powerful computers. What than can we do with this new power? For a starter, let me offer a couple of example of what people has been doing with it:

Retail business intelligence. Retailers are using digital surveillance cameras to capture and analyze customer behavior to generate data that help businesses adjust their operation. One company call this kind of system Visual Business Intelligence (VBI). VBI can for example, figure out how many batch of purchases are being cue up at each registers and avoid the pitfall of counting people instead of groups of people making a purchase together. VBI can also identify difference in purchase pattern (item selection, route around the store) of adults when they are with children and without. I can imagine a future where a more sophisticated VBI can also identify details such as customer emotions by analyzing their facial expression, time they spend looking at product descriptions etc.

Student of Tel Aviv University has a developed a software that identifies facial features, distinguish a female face from a male face and rank attractiveness.

These are just some of the things that exist today, but there are numerous potential for DVA, i have come up with a few my self:

New video game controller. Wii was revolutionary because it allow players to play games in a way that is similar to the action players are trying to do in the game . The Wii remote controller is able to archive this by using a inferred LED sensor bar to sense acceleration along controller’s three axes and where its pointing. Now, Imaging using two high speed cameras instead of the inferred LED sensor to capture the movement of the players hand (may want to get players to put on a pair of uniquely colored gloves). This could mean that even more subtle and detail input by the player become possible, rotations, flick of a finger, a fist, open hand, a punch, a wave. a pinch etc. Of course, it does not have to be a hand, it can be a baseball bat, or a wooden sword, the whole body or whatever is appropriate to the game being played. This kind of technology will revolutionize the console industry more then Wii.

Stealth Bomber detection. Stealth bombers can evade radar detection but they are still visible. So, I guess it would be possible for a system of high speed video cameras and recognition software to scanning the sky for stealth bombers, once identify, it should not be too difficult to know its exact coordinate, speed and direct and subsequently these information to a surface to air missile to shot it down.

Identify fashion trend. What if a software is design to recognize different style of outfit people are wearing in mass numbers base on the video image gather through digital surveillance cameras around the world. That kind of software should be able to identify color coordination (main color, color variation, upper/lower body color variation), types of clothing (one/two piece, jeans, suit, T-shirt, shorts, mini-skirt), pattens (strips, borders, squares) and even material (leather, cotton, hemp, wool). This kind of technology can potentially be used to spot trends. If for example an patten emerges where the most sighted fashion in New York is always follow by major world wide sales of such fashion six months later, this kind of information would be useful to fashion labels.

Really, the potential application for DVA is enormous, its like searching for a beautiful shell on a virgin beach, there are plenty to be found. Companies whom can discover many good ideas very quickly will be positioned to make lots of profit, but how?

Decentralized product strategy

I believe for a technology that has so much potential in such diverse field, it should not be using the traditional centralized product development strategy where a designated group of people in organisation generate ideas, it should be everyone job. How can this be done? Simple, use an online forum where everyone in the organization can participate and post their idea for a product (Hat tip: Linda A. Hill). Unlike a group brainstorming session, where only one person can talk at a time, an online forum allow everyone to express them selfs whenever the idea come to them. And a forum has a kind of wisdom of crowds mechanism build into it, new ideas, popular ideas (by number of responses or views) can quickly be identified, on the first page as the result receive more exposure. This is what i mean by decentralized product strategy. A forum allow people the time to craft their words and others to discuss, refine and improve upon it. However forums can have some minor problems needs addressing: People leaking forum content to competitors, solved by having everyone with access to it sign confidential agreements. Low participation, employees may not want to post ideas or comments that they think are not going to impress co-workers, that can be fixed by enforcing a minimal post per week rule, that encourage people to post their ideas or comments online even if it does not seem clear or refined. Junk post will get filter out by being push back to back pages anyway, so there is no need to worry about forum being flooded by it. Online forums has an unsolvable disadvantage compare to face to face brainstorming session in that some thought are easier to express in person than online, but i think that’s a small drawback compare to the potential of online forum. Once we got loads of ideas generated, how should the company go about commercialising them?


Each of these new product ideas should be developed outside of their original organisation. Instead of building a new department around an new product or assign the new products to existing department, these new ideas should be developed by the people who came up with the idea and is willing to use their own money to develop them into a new independent spin-off business. An arrangement can be made between the original business to own a part and have a sit at the directors board of the new business. Building a new independent business can help avoid a couple of common pitfalls where new product development can be jeopardise by the interest of existing products. One infamous demonstration of such conflict of interest is Sony’s hesitation to develop digital portable music players because it did not want to hurt its music content business by making digital music mainstream for fear of piracy, as a consequence, Apple took over the portable music business. A common threat to many successful tech venture firms after initial public offering (IPO) is many of its core members makes tonnes of money and leave to develop another venture business. Google has and is still losing many of its talent, despite having the best employment environment. Some suggest it is because post-IPO firms loses talents who were attracted to the feel of a venture business. Allowing a product to be its own independent business and let the people who has came up with the idea, most passionate an willing to invest their own money to run it can avoid the two problem mentioned above and still make a health profit for the original company.

To Conclude

The theme of this post is really to promote the idea of decentralization. If there is anything that we have learn from the last century is that decentralization works and centralization don’t in most cases. Although there are rare exceptions where a centralization or a mix of both might work to a limited extent (Steven Job’s Apple, Singapore, Post-Mao China and Japan’s developmental state). This is because other than an abstract direction, how technology develop beyond the frontier is inherently unclear. It is precisely this vagueness that require decentralized agents all independently searching for a path forward. A whole book can be written on decentralized strategy, and I of course do not claim to know anyway near enough about it. I will later attempt to conjure up a post on it, but for now, this post ends here.


One thought on “Potential applications for Digital Video Analysis and a decentralised business strategy to develop it

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s