Ideas for Improving Picasa Web Albums

Google have been doing some promotion of Picasa Web Albums via a "tip" that appears at the top of search queries containing certain trigger words.  Specifically a search that contains the words "photo" and "sharing" together in that order gives a tip of

Tip: Want to share pictures? Try Google's Picasa Web Albums

Searches that just contain "photo" or similar words give a tip of

Tip: Looking for pictures? Try Google Images

Similarly searches related to blogs and blogging have a tip directing users to Blogger.

 

This behaviour has been criticised by a number of bloggers, who view it as human tampering with Google's search results.  In the middle of such a discussion on the Google Blogoscoped forum, there was an interesting exchange between Philipp (the site owner), and Ionut, a regular contributor:

Ionut asked:

Philipp, let's say you are Larry Page, you have a product called Picasa Web Albums, with a market share of 0.04%, and you want more users. What would you do?

Philipp replied:

I'm billionaire then, right? So I'd probably take a loong vacation first. :)
Seriously, I would:
- rename Picasa Web Albums to Google Pictures (or Google Photos)
- move it to pics.google.com
- make it possible to upload PNGs, BMP, TIFF, PCX, etc., which will all convert to JPG, PNG or GIF for display (but the original version will keep intact)
- put some unobtrusive AdSense in the application
- offer the option to have real password-protection/ invitation system for private pictures
- easily allow people to add a Creative Commons license to their pictures
- advertise on Flickr by buying Flickr ads (if Flickr allows this)
- advertise on blogs using the various ad networks/ AdWords-AdSense
- put a link to Google Pictures onto the Google homepage for a couple of days, to see how people react (the spot below the search box)
- potentially, put Google Pictures in the homepage's "more box"
- create a clearer navigation for Google Pics because the current one is partly confusing
- make it easier to understand what others visiting your URL will see; this is completely confusing right now
- add some killer features to Google Pics, like pattern matching search algos, that will draw many people to the site and increase word-of-mouth propaganda among tech-savvy people (who will then evangelize the not-so-savvy crowd)
- create a Google Pics-wide search engine for all photos, not just favorited albums
- allow much more storage; beat competition like Gmail storage beat competition when it was released
- create an online photo retouching tool using the latest technologies; integrate this with Google Pics
- additionally, buy Corel PhotoPaint, fix the bugs, and release it for free – and add an upload functionality to Google Pics
- integrate Hello functionality, kill Hello (but announce this in a blog post first giving transparent reasons)

In other words, I'd do what every other competitor with an image app will have to do to gain market share: improve the product. Only monopolies stop improving the product, which in the end, is bad for the user.

Would I advertise the product as a "tip" in search results? Only if that wouldn't go against what I previously told people. Only if that wouldn't go against core beliefs previously stated. Whether or not the current tip does that, again, everyone can make up their mind about. I consider it a tipvertisement, so it's neither a clear advertisement nor a clear tip.

Ionut provided his own suggestions:

Here's what I'd do.
- Integration with Gmail. You can save photo attachments to Picasa Web and send photos from Picasa Web.
- Integration with Image Search. Everytime you perform a search, you can see search results from your photos / your contacts' photos.
- Move Picasa online. Photo editing, effects, collages.
- Create an extension for Firefox to upload photos faster.
- Upload / view photos on mobile phones.
- 5 GB space storage for free.
- Add NevenVision technology for object recognition, face recognition.
- Every photo you upload to Google in each and every service (Blogger,Docs) should also be saved in Picasa Web.
- Easy geo-tagging, integration with Google Maps.
- How to make money:
* backup CDs
* high-quality prints
* find people from your photos in other public photos

These suggestions are not particularly new, just it's useful to see them all together.  The need for a full search feature in particular is one I've highlighted a number of times before.

One item that's not on these lists is to start official blogs for Picasa and Picasa Web Albums - there's plenty to talk about for these two Google services, and it's silly not to have an official outlet for them.

I think a lot of the under promotion of Picasa and Picasa Web Albums is due to the small, still largely independant team that work on these.  Perhaps Google should look to increase the size of the team significantly - it's unclear yet whether the Neven Vision staff have been added to the Picasa team, or have ended up elsewhere within Google.

Like I care about shoes...

Picasa enables me to find my photos on my own PC, but a lot of the power in that search comes about because of the good metadata associated with the images - I carefully file them in directories where the name means something, and by virtue of the date being stored with images I can often narrow the search down by knowing around when I took a photo (or more likely, whether it was before or after another photo).

On the web however, image searching is a different matter.  Metadata is much harder to find - images prepared for the web often have dates and EXIF metadata stripped off or altered, and images are not filed in neatly named categories.  Most search engines therefore rely on just approximate clues about the image - the filename, link text, other on page clues.  The result is that image search is a very poor relation to text search in many respects.

It's worth noting that there are a number of motivations for doing an image search:

• to find an image of a particular item - this may be the obvious one, but its actually not one that drives most of my searches
• to return to a website that I remember was illustrated by a particular image
• as a sort of visual snippet to help decide which sites are worth visiting (on the basis that if a site goes to the trouble of sourcing good images, the rest of its content may be of a similar standard)

Thus around a year ago when Riya launched, I was hoping for a site that would push the boundaries of image search.  Their particular hook was that they did textual and facial recognition.  The facial recognition got the big press, since it was the sexier technology, but was beset with problems - it didn't work well enough to live up to the hype, and had it done so, there are all sorts of privacy concerns, with shades of Big Brother being able to find lots of images of an individual automatically.

I was therefore more interested in the textual recognition that they offered.  However, if possible, this was even worse than the facial recognition.  The recognition rate was quite low - many images showing text had nothing in them recognized, and even when text was recognized, it was normally as single words, rather than as a block of text.  (My understanding from Tara Hunt is that this was not a technical issue, rather that the licence to the technology that Riya obtained to do this would not permit them to do continuous text recognition).  There are a few images that the text recognition has tagged - for example search for "exit", but in general so little text from images was recognized that you probably forget the site ever offered this feature.

Riya realized they had problems, (their CEO writes a very open blog on the matter), and decided to change their emphasis, to become a real visual search engine.  This sounded great, and I eagerly awaited the arrival of Riya 2.0 as they considered it.

And today it arrived - and I don't like it.

Rather than perfecting their technology for a wide ranging search field, they have instead produced a shopping service - you can use it to search for products that look like each other - provided that those products are shoes, watches, handbags or jewelry.  I'm certainly not in their target market - 20 to 30 year old women, so perhaps I'm missing the point, but these seem a strange set of goods to handle.  Shoes are one of the last things I would buy on-line - unlike most clothes they need to fit pretty exactly, so generally need to be tried on in person.  Watches and jewelry are high value items - and not everyday purchases, so not an obvious online purchase either.  And as a man, I'm never likely to understands handbags - so again the exact logic behind offering this product escapes me.  Perhaps the reasons for these choices lies with the celebrity pictures to be found on the site - where the pictures show watches and shoes highlighted, and linked to searches.

All in all, with the widespread coverage of like.com's launch, it does look as if I'm not alone in thinking they could have done more with this technology.  Here's hoping that Picasa is able to do more with the related object recognition technology they recently got via Neven Vision.

Good further coverage of Like is to be found at: Robert Scoble, Thomas Hawk, ZDNet, TechCrunch, Greg Linden, GigaOM, Don Dodge.

It's worth noting that visual search for shoes is not unique to Like - Chez Imelda also offers this, powered by technology from Pixsta. 

JotSpot Photo Gallery

And today's acquisition by Google, is JotSpot, following their recent purchases of YouTube and Neven Vision.

JotSpot is a hosting platform for providing wikis, but it goes beyond the normal range of (mostly) text wikis, by offering a rich application platform, including a number of preprogrammed web applications that can be included within your JotSpot wiki.

One of these is the JotSpot photo gallery.

This offers the ability to

• Create albums and upload photos
• Thumbnail and slideshow views of your photo albums
• Comment on photos when they are uploaded
• View photo info (EXIF) for uploaded photos
• View a photo stream of recently added photos across all albums
• Read a comment stream of recently added comments across all albums

Hardly earth shattering I know, but the very fact that that list is so familiar is indicative of the fact that photo galleries are an intrinsic part of the wide collaborative platform that the web is becoming.  The ability to comment on photos in a gallery, and to get streams of information when more comments or photos are added is a very useful feature.

Picasa Web Albums offers these features of course, and I don't see that the JotSpot Photo Gallery will replace the Picasa Web, but the obvious "fit" for JotSpot's collaborative approach is within the "Documents and Spreadsheets" package that Google already offer.  Picasa Web has a foothold there already, appearing as a "photos" link within the spreadsheets and other documents, and I think that this move is likely to help make this a closer integration.  At the moment you can add just a short caption to the photos appearing in a Picasa Web Album, but close integration could see photos appearing as part of much richer collaborative documents - as cells in spreadsheet for example, or embedded within paragraphs of text in a word processor document.

Even just the ability to add more to the photo album pages themselves within Picasa Web would be immensely useful - I've seen a good number of requests by users to add counters or analytics code to the pages, and JotSpot is all about the ability to easily change page content, so would provide a useful model to follow.

Google makes acquisitions for a number or reasons - including to acquire technology, and to acquire people.  JotSpot certainly has some bright people, and although the wiki technology may be far from unique, it was packaged well.  I look forward to seeing the wiki ideals of easy collaboration making their presence more widespread in other Google properties, and Picasa Web looks a prime candidate for many of these ideas to show up first.

There's a number of posts from Google about this - the main Google Blog carries the announcement of the deal and a follow up that notes how Google have been pushing the collaborative angle for a while already, moving  "calendars, photos and documents onto the web".  Of course the JotSpot blog now also becomes another Google blog as well, and Ken Norton's personal blog also relates the news.

Some graphic skills that could do wonders to Picasa

Via an entry on the Google Code blog, Announcing OpenVis3d, which pointed out that Google is "very interested in computer vision and graphics", I was led on to this Google job opening for a Software Engineer, Computer Vision and Graphics.

There's an impressive list of technologies listed there, and although I rather think this is position is about Google Earth (the clue is perhaps in the question "would you like to build the largest image ever, a petapixel multi-resolution 3D mosaic of the earth using a few thousand computers and millions of source images?"), more than a few of them could be applied to Picasa with good effect.

The list includes

• image stitching - [yes please, it would be great if Picasa could help form panoramics and other stiched images easily]

and a whole load of techniques to help understanding what the picture might contain, to make searching images more precise

• structure from motion or shading
• object detection
• document image understanding
• pattern recognition

Picasa has a couple of "experimental" menu options that allow for searching based somewhat on the content of the images - the search by colour, and the "find duplicates" options, but these are fairly easly attributes of an image to work out.  Deeper understanding of what an image contains, such as objects or text, will add considerably to the search abilities of the program, and indeed at the time of the Neven Vision acquisition, the stated aim was that this facial and object recognition technology would find a place within Picasa.

Google buys YouTube

There has been masses of comment on the Google purchase of YouTube for $1.65 million in stock.  There is some thought provoking stuff at Geeking With Greg, to choose just one site of many.

Another well thought out piece is Facial Recognition in Videos which considers:

I wonder, however, if part of Google’s technology strategy is to find a way to index and search videos in a whole new way. Google acquired Neven Vision without too much fanfare in August. Neven Vision develops face recognition technology and Google presumably has plans to integrate face / person recognition into Picasa, their photo management and editing application.

...

Could it be that Google has also recognized the potential of a true visual search for information - or pictures - or videos. Could Google have plans to enhance the Neven Vision image recognition technology to apply it to video?

That's a interesting question, and I'm certainly very keen to see what the Neven Vision technology can bring to Picasa and other Google properties.  Video search is hard - even more so than images, since the video may cover many subjects throughout its length, so automatically gathered metadata may not be particularly targeted.  The tags on YouTube go some way to addressing that - and by now controlling the largest chunk of user generated video content available, Google have access to a huge training set for their automatic recognition programs.

Google Image Tagger

Google has been registering some new domain names, and amongst those are googleimagetagger.com/net/org.

For the moment we can but speculate what sort of service this might be - whether the images are tagged automatically (such as face and object recognition via Neven Vision provided technology?), or manually as in the widespread tagging at flickr. Up to now, Google has tended to avoid "tagging" at least in name - though Gmail labels are close, and Picasa's keywords share some similarities.

Neven Vision adwords campaign still running

With the flurry of blog posts about Neven Vision, there are now loads of pages whose content matches up with Neven Vision's adwords campaign.

Despite Google's efforts to wipe Neven Vision off the web (clearing its main and Japanese websites completely), they seem to have missed the fact that the adwords campaign is still running...

The ad I've seem most is

Fastest Face Recognition
and Facial Feature Tracking SDKs and Custom Solutions
www.nevenvision.com

which points to the (now non-existent) page www.nevenvision.com/biometric.html

Visual searching via mobile phone

One of Neven Vision's previous investors was the venture capital firm Zone Ventures, who generally provide $1M - $3M level of funding.

Via the links page they maintain on their portfolio companies, there is this link to a video from June 22, 2006 titled "Neven Vision - The Next Big Thing in Merchandising!"

It's a short (2min 40sec) clip from an NBC4.TV news program, showing Neven Visions mobile phone search application in action.

• Take a photo of a CD on your mobile phone camera, and link to the groups latest video, or to the album lyrics
• A photo of a Coca-Cola can links you to info on the 2008 Olympic Games (which Coke sponsor)
• A photo of a Starbucks cup gets you a coupon
• A photo of a movie poster links you to the relevant website
• Consumer electronics, link you to reviews of those products
• A car, allows you to access blue book values
• ... and perhaps a photo of a person, to get info on that person.

"Neven Vision's first US campaign using this technology will be for a car maker. It debuts this fall."

Riya and Google - from the inside

Many commentators have picked up on the hidden story in the Neven Vision acquisition - that this rather closes the door on a Google / Riya hookup.

Munjal, the Riya CEO, has responded to the questions with a second blog post.  He states that he is under NDA, so can't say all he wants to (but will be checking as to when this expires, so that he can tell all then).

In the meantime, he gives a timeline of

a) In Nov'05 you heard a lot of rumors about Google and Riya
b) In January'06 you heard me say that we are still an independent company
c) In January'06 you heard Riya announce it just raised a large series B financing of $15.5M - large rounds are usually pre-cursors to fund large new bold (and risky) strategies.
d) In March'06 Riya launches Riya 1.0
e) In May'06 Riya launches an expansion of it's vision to Riya 2.0 Visual search for the web = bold big strategies
f) In Aug'06 you hear about Google buying Neven Vision to deliver Riya 1.0 functionality to Picasa

He then goes on to discuss what this means to Riya going forward.  He makes 3 points

1. Google is often not the best of breed solution outside of its core search competancy
2. Adding face recognition to Picasa adds a feature that Riya have now all but abandoned anyway
3. Now that Google is openly in competition with Riya, the partnership potential at Riya with other firms is looking increasingly rosy.

It's a positive message, but I think he may be missing the point.  I don't think that Google bought Neven Vision just to allow Picasa to catch up with Riya 1.0 (client side face recognition, with upload ability).

I think it more than likely that Google will be introducing "machine vision" (not just face recognition) driven features across far more of their services than this.  This puts them squarely in competition with the refocused Riya 2.0 - which is now positioning itself to be a visual search engine, based on similarity of images.  With the expansion of the Picasa brand into web galleries at Picasa Web Albums, images are just as likely to be subject to machine vision on the server as they are on the client.

In addition, Neven Vision has a strong presence in the mobile phone market (and the patent portfolio to back it up), and Google is increasingly making strong moves in the mobile data market.

Public information on Neven Vision - disappearing fast

One of the problems with the web is that information is often ephemeral, (as distinct from information in say books that remain available via libraries for perhaps hundreds of years).  Despite its mission to "make all the world's information accessible", and the strides that Google Books is making in unlocking some of the information locked up in libraries, Google itself is notoriously bad at preserving the information it is personally responsible for.  I've spoken before how it has withdrawn masses of useful information in its own products forums, and we are seeing the pattern repeated once again with its new acquisition, Neven Vision.

When I first posted yesterday on the Neven Vision sale, the main website was down.  It came back up later in the day, allowing a lot of background info on the company to be read, but by the end of the day it had all been wiped. Now the site merely contains the simple announcement: "Thank you for your interest. Neven Vision was recently acquired by Google Inc. and Neven Vision product information is no longer available on this site. Click here to learn more."

In an effort to preserve some of the knowledge that the site imparted, here are a few gleanings from notes I made earlier:

---

Products cover two areas - Biometric Identification, and Consumer Applications.

Products for consumer applications include

• iScout, which allows camera phone users to take photos to initiate a search process and get relevant information or content sent to their phone. 
• Photo management for phone handsets - where face recognition adds automatic tags to allow sorting of images by who they show
• Mobile security, where the phone handset only permits users whose faces it recognizes to do certain things, such as access email
• Masquerade - to add accessories such as hats, masks and glasses to photos
• Delegate - uses face detection and facial feature tracking to drive 3D avatars for enhanced chatting and video conferencing from a mobile camera
• Image search - facial matching to find friends wherever they may have posted their images, and to spot similarities to celebrities

Partners for consumer applications include Coca Cola, Jamster, NEC, Samsung, vodafone, Toshiba, Sharp, Logitech, NTT DoCoMo.

Biometric Identification products are used by a number of law enforcement agencies.

---

Summary of the technology

Computer Vision, also known as Machine Vision, is a field of technology engaged in teaching computers to "see" and understand visual imagery. Neven Vision is primarily focused on the subset of computer vision known as visual sensing , which is the automated extraction of information about objects or scenes in one or more images.

Neven Vision's computer vision technology is the fastest in the industry and represents leading accomplishments in this field. Several U.S. patents have been granted to Neven Vision for its unique inventions. U.S. government and independent observers have recognized the underlying technology as most advanced with respect to accuracy, speed and efficiency.

The core building blocks of the company's vision-enabled products and services are our patented face and object recognition engines. Below is a simple representation of how each of these engines work.

Face Recognition

1. Faces in an image are automatically detected by a robust face finder component that determines the position and size of the face in real world conditions (varying illumination, pose and expression).
2. A second processing stage determines the position of local features on the face, called “landmark finding”.
3. Then a Gabor Wavelet transformation takes place to compute the “face-template” from the local features by extracting the template from an image that contains the essential, condensed facial information needed to determine a person's identity (a “face-template” is only about 1kByte in size; min <350bytes max 1.6KBytes). Two templates are then compared to yield a similarity. Templates belonging to the same person produce high similarity while templates from different persons produce low similarities.

Object Recognition

Neven Vision's object recognition technology is based on a local feature approach:

1. Objects are recognized by identifying characteristic points of an object, deriving feature vectors from the texture in the vicinity of the characteristic points.
2. These feature vectors are then compared to a database of known objects to establish matching correspondences between the current object and the objects contained in the database.
3. If a sufficient number of local feature correspondences is found, a positive identification is accepted.

---

Office Locations

LOS ANGELES, CALIFORNIA
Neven Vision
2400 Broadway, Suite 240
Santa Monica, CA 90404
PHONE
(877) 6-NVISION
(310) 828.0898
FAX
(310) 828.0479

TOKYO, JAPAN
DAI2 Okamotoya Bldg. 4F
1-22-16, Toranomon, Minato-Ku,
Tokyo 105-0001, Japan
PHONE
(81) 3-5251-5631
FAX
(81) 3-5251-5632

HAMBURG, GERMANY
Neven Vision Germany GmbH
Ziegelweg 5
D-26188 Edewecht
Germany
PHONE
+49 180 33 48899
FAX
+49 180 30 03331 2914

MUNICH, GERMANY
PHONE
+49 89 89 99 99 77
FAX
+49 89 89 99 99 78

UK
PHONE
+44 (0) 7956-264-27

---

Executive Management

Alex Cory CEO
Dr. Hartmut Neven Founder AND CTO
Jordan Posell CFO

BIOMETRICS

Piet Lesage General Manager, Biometrics
Catharine Evans  Director, Biometrics

TECHNICAL & PRODUCT MANAGEMENT

Detlev Schwabe VP Product Engineering
Ameen Ahmad  Sr. Director, Product Mgmt & Mktg

Hartwig Adam  VP Platform Technology

SALES

Stefan Fleissner  VP, Europe
Craig Patton GM, UK Operations
Paul Cushman  VP, US

Axel Boesche GM, Neven Vision Germany
Bret Crochet VP Bus. Dev., Southeast Asia

FINANCE

Steven M. Kantor, CPA VP Finance

BOARD OF DIRECTORS

William Woodward Managing Director, Anthem Venture Partners
Dr. Hartmut Neven CTO, Neven Vision
Niloo Howe Partner, Paladin Capital Group

Todd Jerry Partner, Anthem Venture Partners
Barak Bussel Partner. Third Wave Ventures

BOARD OF ADVISORS

Dr. Roland Deiser Managing Partner, Kingstone Partners
Jaron Lanier VISIONARY
Hiroshi Shin Ohashi CEO, Big Bridge

---

List of patents

• EP1072018 Wavelet-Based Facial Motion Capture for Avatar Animation
• 1072014 Face Recognition from Video Images
• EP1072018 Wavelet-Based Facial Motion Capture for Avatar Animation
• 218457 Face Recognition from Video Images
• 218458 Wavelet-Based Facial Motion Capture for Avatar Animation
• EP1072018 Wavelet-Based Facial Motion Capture for Avatar Animation
• 1072014 Face Recognition from Video Images
• 6714661 Method & System for Customizing Facial Feature Tracking Using Precise Landmark
• 6222939 Labeled Bunch Graphs for Image Analysis (EYEM1160/ NE01)
• 6356659 Labeled Bunch Graphs for Image Analysis
• 6563950 Labeled Bunch Graphs for Image Analysis
• 6466695 Procedure for Automatic Analysis of Images & Image Sequences Based on Two Dimensional Shape Primitives
• 6272231 Wavelet-Based Facial Motion Capture for Avatar Animation
• 6580811 Wavelet-Based Facial Motion Capture for Avatar Animation
• 6301370 Face Recognition from Video Imag

Face recognition coming to Picasa - Google purchases Neven Vision

It looks as if face recognition is on its way to Picasa after Google announce on their blog that the Neven Vision team is now part of Google.

In announcing the move, Adrian Graham, Picasa Product Manager, notes that:

It's not always easy to search through your personal photos, and it's certainly a lot harder than searching the web. Unless you take the time to label and organize all your pictures (and I'll freely admit that I don't), chances are it can be pretty hard to find that photo you just know is hidden somewhere deep inside your computer.

We've been working to make Picasa (Google's free photo-organizing software) even better when it comes to searching for your own photos—to make finding them be as easy as finding stuff on the web. Luckily we've found some people who share this goal, and are excited that the Neven Vision team is now part of Google.

Neven Vision comes to Google with deep technology and expertise around automatically extracting information from a photo. It could be as simple as detecting whether or not a photo contains a person, or, one day, as complex as recognizing people, places, and objects. This technology just may make it a lot easier for you to organize and find the photos you care about. We don't have any specific features to show off today, but we're looking forward to having more to share with you soon.

Neven Vision had offices in Los Angeles, Tokyo, Munich and Kuala Lumpur. It's unclear whether the team will all be moving to the Picasa offices (also in LA).

The Neven Vision site is unavailable just now slick to look at, but the downloads/demos are listed as "coming soon"! The Japanese operation site N-Vision is still up and running.

The team which grew out of university research in Germany are very well regarded for their machine vision technology - it's not just facial analysis, but also more general object analysis. The company was founded by Dr Hartmut Neven in 2003, and capitalized at 140million Yen, which is about US $1.2 Million.

Update: Plenty of comment on this across the web:

• Google acquires Neven Vision by Munjal from Riya.com (which also uses facial recognition).
• Google May Soon Tell You Where Your Picture Was Taken
• Object Recognition Is The Future Of Google
• Google Buys Neven Vision
• Google acquires biometric company Neven Vision

Update:

William Slawski has done some good research into the many patents that Neven Vision hold.