How Picture Recognition Leads to More Revenue Recognition

Nstein recently announced a technological partnership with Imprezzeo, an image-based search company located in London and Sidney (cool place BTW) 

I have to say that when I first saw their standard demo a couple of months ago, I was immediately seduced by the potential of their technology tied to our products. Thanks to a smart and lightweight integration in our DAM and WCM, there is no doubt that Nstein can provide users a quicker and more relevant image search experience by using images to search for images rather than text.

 

As a matter of fact, the Imprezzeo software is based on a unique combination of content-based image retrieval (CBIR) and facial recognition (FR) technologies that bring immediate benefits to our customers. Let me explain.

 

Today, thanks to our text-mining engine (TME) Nstein already provides exclusive processing of XMP metadata and analysis of photo captions to best leverage the text describing the images. The title and captions are linguistically processed with specific algorithms in order to take most benefits of that information by extracting hot topics, relevant concepts and categorizing photos against dedicated taxonomies.

 

However, that’s where we are reaching the limit of the technology… Why? Because, unfortunately only a small percentage of pictures contains relevant metadata — and the amount of keywords required to really describe all the facets of a picture is just too big.

 

Therefore, pictures coming from news agencies of photo stock companies are more or less well tagged but the huge majority of pictures coming from photographers or from direct channels (newspapers and magazines receive more and more pictures coming from MMS and email) are generally poorly (if at all) tagged. Nevermind that captions may be written in other languages!

 

Thus, the best-tagged pictures are generally much more expensive than the non-tagged pictures, which may even be free but unless iconographers enrich these assets manually the classical text-based search will just fail…

 

That’s where Nstein + Imprezzeo = Benefits, or in other words; the whole is greater than the sum of its parts.

 

The list of immediate benefits brought by this partnership include:

• New search experience by first leveraging existing captions and text-mining to initiate a semantic-based image search and then drilling-down/refining the search using a “freestyle browsing” approach based on similarity and image recognition (better user experience, navigation by image vs. text, content discovery)
• Automation and acceleration of tagging and categorization processes by gathering all similar pictures and enrich them in a single operation (increase performance and efficiency)
• Ability to choose the least expensive image among a set of very similar pictures (direct costs saving!)
• Finding “hidden gems” by discovery methods or “more like this” feature 
• Ability to monetize more assets that publishers already own (new source of revenues)
• Ability to provide syndication partners with more choices!

 

I am looking forward the integration to come… very promising! I will keep you posted.

A Successful REST Implementation Brings Benefits To Customers And Software Vendors

Not a surprise, and since a couple of years now, the RESTful architecture style promoted by Roy Fielding is more and more adopted by the community… and that make sense specifically in the domain of web-based applications and web-distributed applications. I will not emphasize the importance of REST in this post as (too?) many contributions have already been published by various experts.

However, for software companies who has already a range of existing products and solutions it is not always easy to bend their development roadmap to encompass the requirements of this architectural style. In fact the consequences and let’s name them, the “collateral damages” can be significant. 

Although, REST for REST can be pointless if not supported by a real strategy and business benefits. At Nstein, our core offer is composed of 2 products (a web content management software and a digital asset management software) which leverage a powerful text-mining engine (what CMS wire call our ‘secret sauce‘).

And there we go: At the end of 2008 we decided to make sure that our offer will be 100% REST in the 6 months to come… and we did it!

From a management standpoint, I can see several reasons for this apparently easy success that I would like to share with you:

• First of all, WCM and DAM are products which, by nature, manage resource. Therefore a RESTful architectural style was a natural fit.
• The “RESTification” or our solution was a global project involving every actors in the company from the product development and QA to the implementation services, support and even marketing and sales. This global commitment was key to ensure a neat change management.
• REST was not the objective but the means! The means to improve our scalability and ease of deployment of course, but also another way to reinforce our global “user-centric” strategy.
• The R&D organization has been adjusted to better match this new challenge and ensure the most reactive but still effective work.

Actually, for this last two points (REST as means to support our objectives and new organization) we decided 5 month ago to create a dedicated UX/UI team in the R&D department to focus our efforts on usability and user experience in general.

And there you are: in the same way that you cannot expect from chiropodists to be experts in microsurgery (even if both of them are doctors) you should not expect that a programmer can be an expert in mathematics, database optimization, Java, C#, PHP, Python, Ruby, UI design, XHTML, CSS development, and can even fix your computer! (even if sometimes my Neighbours or my Parents do believe that ;-)

This to say that IT people has different skills and poles of interests. However, most of those skills are required to release a whole user-friendly product. So, by intentionally splitting the responsibilities and development tasks between a “hard-core” team and an “UX/UI” team (and the REST design was the perfect excuse) we have focus on the best skills of each person.

The effect is that we have reached our global objectives in less than 4 months with quite satisfying benefits: As geeks were enjoying focusing on tuning, performance and scalability, the designers and UX/UI programmers were enjoying creating cool and dynamic AJAX-based interfaces…

The development team itself has moved from a product-based approach to a competence centers organization that not only fully leverages the talents but enables a real agility (as you have guess, ‘Agile‘ is the development methodology we use and promote)

Thus, thanks to a combination of REST style, organization changes, strategy, willingness and several lines of code, Nstein provides a better than ever user experience while guarantying that our solutions will stay innovative, efficient and scalable.

So if you are designing web-based products and applications, do not rest until you are RESTful ;-) but be sure to include your architectural choice in a global strategy supported by those – too often forgotten - business benefits!