Riya 2.0 Roadmap
For those of you who came to this post from our homepage: Welcome to my blog. I am Munjal Shah, co-founder & CEO of Riya.
For the rest of you welcome back.
Change is afoot. At this point our new site has gone out (or should shortly..;-)
We have all been working our tails off getting Riya 2.0 ready (read my description of Riya 2.0 here and read Matt Marshall's thoughts here ).
The site you see today is our first of four releases, and is a very tiny step toward delivering a web wide visual search engine.
In the next few months (3-4) we will roll Riya 2.0 out slowly. Why roll it out piece by piece? As all engineers know, it is easier to test and deploy small changes than to do it all at once. Eventhough a single big release is great from a marketing standpoint, we are more interested in getting this out and learning from your feedback as fast as possible. Why not be secretive and not show where we are going? We just believe inclusiveness and transparency beat paranoia and fear everytime.
Note: In this release we don't have similarity turned on yet.
This release just lays the ground work by putting in place a lot of architecture and UI. So for now you'll have to settle for screen shots... sorry. Below is a preview of what our Visual Search will eventually allow.
This is what the new homepage will eventually look like. We recently had a homepage competition which Jeff and Andrew Hitchcock won. I think you can probably see how their color wheel search inspired this page. It isn't exactly what they designed but their key break-through was allowing people to search not just using text. We expanded on this idea considerably by adding the search by sketch idea which Retrievr implemented with a small demo of 70K images. In addition, we added search by camera phone and upload. Once this is out you'll be able to take a photo of a rug you like at a store and then find similar rugs on the web.
We'll also let you find similar images from a text search by clicking on the "More like this button" on the right of the photo preview (the preview is a new feature that is already live - try it out and let us know).
Below are some other screen shots of what Riya 2.0 will look like and some ideas of similarity searches across the web that you'll be able to do. BTW this UI was mostly done my Uma in our Bangalore office. I think he did some great designs here.
Suffice it to say we are really really excited.
That being said, it has been a tough couple of months. Getting the whole company reorganized around this new larger vision took its toll on all of us. I must say there are many times we felt like we jumped out of plane with cloth, tread, and a needle and are sewing a parachute in mid-flight. I guess that is what it is to build a startup...;-). Azhar, Burak, and I are frankly more stressed we've ever been.
At the same time, our new strategy is so simple and powerful (if it works).
Baris and Vincent are driving our face and image similarity teams on the research side and doing a great job. Nikhil, Sandeep, Amod, Neelesh, Tanvir, and Dan have been driving the engineering side. You can just feel the freight train of releases in Riya picking up speed and bit by bit implementing Riya 2.0.
We are all holding our breath and hoping that once this is out - you all will love it and use it all the time. Until then we keep working away.... Over the next few months you'll see
a) The number of photos we index grow from the relatively small number today
b) The face similarity and image similarity come online
c) The ability to upload any photo come online
d) Much more...
For those of you use our Riya 1.0 product. It is still there. Click on the "My Photos" tab and login. We've made some improvements to it as well.
Thanks for listening. Now please write us back at feedback2 at Riya.com and we'll listen.










That looks cool. I'm excited for the launch and playing around with new features.
Posted by: Andrew Hitchcock | July 28, 2006 at 02:44 PM
Hi Munjal,
I've been excitedly following your account of Riya's first 60 days of life and since the last episode isn't out yet (not that I'm gently cajoling you here to finish it :), I'm thanking you right here for it. I know it must have been a big draw of time and effort for you to do it and considering your tight situation as CEO of a major web2.0 startup it's all the more admirable. Thanks, I had much to learn from it.
Anyway, I wanted to give you something back in return and since I'm poor, I could only muster my 2 cents. I wrote a (pretty) long essay on Riya to which I gave much thought: http://www.elzr.com/articles/2006/07/31/an-essay-on-riya. I hope you're able to read it and would be happiest if it were somehow profitable to you--you've given me too much already.
Thanks again for your blogging and good luck.
Posted by: Eliazar | July 31, 2006 at 01:03 PM
I can't wait to see this working!
Posted by: art-one | August 01, 2006 at 06:49 AM
I hope this works out for you guys, but I am not 100% sure why you need such a drastic shift from your initial strategy.
It would seem to me that licensing your technology would just be an easier direction. One idea is to license face recognition to facebook.com or myspace, so that if a user has pictures up and it recognizes a face in someone elses albums, it can suggest that you guys know mutual people and try to form a connection...just a thought...
Posted by: Eli | August 01, 2006 at 06:44 PM
well, it would be awesome if u guys added a remember me feature for signups...
its too boring to type the password everytime i login....
it doesnt show up in safari for sure.. if u have done it already
Posted by: madhu Venkatesan | August 05, 2006 at 01:42 PM
I really love this idea. I'm really curious to see where you take it.
Posted by: Joe Mordetsky | August 07, 2006 at 01:15 PM
I was just giving this some more thought, I really love the concept "Once this is out you'll be able to take a photo of a rug you like at a store and then find similar rugs on the web."
That would make a brilliant service and will bridge the gap to monetization.
I guess my one question is what do you see as the hurdles to getting there? My first thoughts are performance. The cold upload of an image and comparison against all others in an index or db seems like a very costly undertaking. I've written a local spider like this that was designed to find similar images using a third party comparison engine.
While the image comparisons were pretty quick for smaller images, the operation itself still took a good deal of time, certainly more then a individual user might sit around to wait for a price on their rug.
Do you guys have technology that elliminate this?
Posted by: Joe Mordetsky | August 07, 2006 at 01:31 PM
Hi there.
I saw the site mentioned in the September isue of Discover. Are you, by any chance, using MPEG 7 methods for characterizing and indexing multimedia content? If not, can your methods be cross referenced?
And do you plan to offer comparable sound search capability?
Cheers,
**Jeff**
Posted by: Jeff Sutter | August 11, 2006 at 09:37 AM
that's darn impressive.
keep up the great work :)
Posted by: Dave | August 15, 2006 at 07:37 AM
Love this idea, have been thinking of this for years, it is major missing component in the Google arsenal or tools.
I think the tech is great, I hope you guys are able to reach the entire picture silos of all the major image datacenters like flickr, etc..
Another idea.. "find a person [using an image] by geographic region", using a recent picture, could this be a Google maps mashup?
anyhoo.. lovin' the idea, may success follow your effort!
Fabian.
Posted by: Fabian | August 15, 2006 at 10:08 AM
Any plans to release a uploader for Linux? Or atleast a web based interface. Because there is no way I can use it now.
Posted by: nilesh | August 29, 2006 at 01:44 PM
I happened to write a related post, doesnt seem to appear in the trackbacks. You can find the post at this link
http://www.venukb.com/blog/2006/09/03/image-search/
I am sure Riya 2.0 will be a very powerful search app, but how do you plan to handle the tagging of images ? I guess Riya also has to take up an exercise similar to Google Image labler.
Posted by: Venu | September 05, 2006 at 09:37 PM
Riya may be one of the tools that revolutionizes how we use the net as we have come to know it. Your ideas and vision for the future of finding and sharing information is truly exceptional and worthy of the archives at www.thinking-forward.com. Thank you for sharing your brilliance with the world.
Posted by: joe bruzzese | September 12, 2006 at 11:19 PM
I'm just curious if this program can match people in older photos, to tell you if it's the same person or not...
Thanks,
Cate
Posted by: Cathy | October 10, 2006 at 08:51 AM
Its DAMN COOL !!!
Posted by: Sanjeet P. Arun | October 11, 2006 at 11:20 PM
I am sure that riya can outwit any image search engines with the help of an AI based engine
Posted by: Sanjeet P. Arun | October 11, 2006 at 11:24 PM
I hope that riya is the name of Mr.Munjal's child
Posted by: Sanjeet P. Arun | October 11, 2006 at 11:28 PM
Riya and your team sure seem to promise a lot of potential.
My wish is that you would do more exciting things for the user base as you move along... continue the buzz of what the technology is doing while you guys can get some proper kudos along the way. I'd love to be able to do a simple genealogy photo organization for instance. Or a simple address book with all the faces per each entry, etc.
There are so many cool apps you could tune your engine to that would pay off in so much excitement.
Have you considered using the Google model where their programmers get a certain percentage of time for personal R&D projects? I think that can yield a lot of fun products that turn out to be profitable in the end.
Best wishes.
Posted by: david sutherland | October 12, 2006 at 12:12 AM
When is the upload-image search feature coming?
Posted by: jay | October 23, 2006 at 10:10 PM
software engineer.
Posted by: sudhir prasad | October 25, 2006 at 10:52 PM
Oh boy, in an almost 3-months-old comment to this post ( http://munjal.typepad.com/recognizing_deven/2006/07/riya_20_roadmap.html#comment-20500518 ) I linked to a long essay --extended thought session-- on Riya that I wrote as a way to try to contribute something to this blog. Today I found out the link was botched: the comment engine on the post turned the plain URL into a link WITH the dot that marked the end of the sentence. Following the link took you to a paltry "Post not found" message. Grr...
I'm sorry. I spent the day trying to get my routing and caching software to understand the URL and it now works even with the dot. Alternatively, here's the URL again, without the dot, as god intended it: http://www.elzr.com/articles/2006/07/31/an-essay-on-riya
Posted by: eliazar | October 26, 2006 at 07:09 PM
Based upon you internet capabilities, i would highly suggest you link onto the website above and see the worlds first and only ALL DIGITAL PLUG AND PLAY LEVEL INTERGRATED VIDEO WEBSITE AND SOFTWARE. Please give me a call afterwards and let me know how you think we can help you in your business to become ONE OF THE FIRST TO OFFER AN ALL VIDEO WEBSITE.
Sincerely,
Wayne Anderton
901-606-1934
Posted by: wayne anderton | October 29, 2006 at 10:23 PM
This looks fantastic. Question though, will you continue with the photo management now that your vision has changed? I'm considering starting to use it more, but if it's going away soon, I don't want to get addicted.
Posted by: kyb | November 08, 2006 at 07:25 AM
Find and delete similar and duplicate images and picturesVisual Similarity Duplicate Image Finder will find all similar and duplicate images in a folder and its sub folders. It uses advanced algorithms in order to find similar images in a way that a human does. Due to that fact it can find similar images even if they are in different image formats and sizes.
http://www.mindgems.com/products/VS-Duplicate-Image-Finder/VSDIF-About.htm
Posted by: Find and delete similar and duplicate images and pictures | May 20, 2008 at 10:35 AM
nice collection
Posted by: website design nyc | May 19, 2009 at 04:48 AM