Kevin Schofield - Tour of Microsoft Research's Next Media group
- Posted: Jul 21, 2004 at 11:36 AM
- 131,889 Views
- 60 Comments
Download
How do I download the videos?
- To download, right click the file type you would like and pick “Save target as…” or “Save link as…”
Why should I download videos from Channel9?
- It's an easy way to save the videos you like locally.
- You can save the videos in order to watch them offline.
- If all you want is to hear the audio, you can download the MP3!
Which version should I choose?
- If you want to view the video on your PC, Xbox or Media Center, download the High Quality WMV file (this is the highest quality version we have available).
- If you'd like a lower bitrate version, to reduce the download time or cost, then choose the Medium Quality WMV file.
- If you have a Zune, WP7, iPhone, iPad, or iPod device, choose the low or medium MP4 file.
- If you just want to hear the audio of the video, choose the MP3 file.
Right click “Save as…”
- Mid Quality WMV (Lo-band, Mobile)
- WMV (WMV Video)
Digital photography. It's one of the hottest product categories. It seems like everyone's getting a digital camera lately, but managing those photos is really hard.
Some of us on the Channel 9 team, for instance, have more than 8,000 digital photos, already (and we know some guys who already have more than 50,000).
So, while we were wandering around Microsoft Research with Kevin Schofield, we wondered "what could be done to make organizing photos easier and nicer?"
Microsoft Research's Next Media group has the answer. Here Curtis Wong, the guy who leads the group, demoed for us their Media Browser technology (News.com called it Photo Triage, but the team actually calls it the "Media Browser" because it does more than photos).
This technology demonstrates how you could organize and look at your digital photos in the future.
Thanks again to Kevin Schofield (and all the researchers we met) for giving us a great look at some of what Microsoft Research is doing.
If you missed the other segments of the tour we got, here they are again:
Interview with Kevin Schofield
Tour of University Relations Group (women in computing)
Tour of Social Software and Hardware Groups (Wallop, MicroMotors, Hardware Lab Tour)
Tour of VIBE Group (large and multiple monitor research)
Kevin has a personal weblog, and you can read more about what Microsoft Research is working on (we only saw a small fraction of what they are working on during our tour) on the official site.
Some of us on the Channel 9 team, for instance, have more than 8,000 digital photos, already (and we know some guys who already have more than 50,000).
So, while we were wandering around Microsoft Research with Kevin Schofield, we wondered "what could be done to make organizing photos easier and nicer?"
Microsoft Research's Next Media group has the answer. Here Curtis Wong, the guy who leads the group, demoed for us their Media Browser technology (News.com called it Photo Triage, but the team actually calls it the "Media Browser" because it does more than photos).
This technology demonstrates how you could organize and look at your digital photos in the future.
Thanks again to Kevin Schofield (and all the researchers we met) for giving us a great look at some of what Microsoft Research is doing.
If you missed the other segments of the tour we got, here they are again:
Interview with Kevin Schofield
Tour of University Relations Group (women in computing)
Tour of Social Software and Hardware Groups (Wallop, MicroMotors, Hardware Lab Tour)
Tour of VIBE Group (large and multiple monitor research)
Kevin has a personal weblog, and you can read more about what Microsoft Research is working on (we only saw a small fraction of what they are working on during our tour) on the official site.
Comments Closed
Comments have been closed since this content was published more than 30 days ago, but if you'd like to continue the conversation,
please create a new thread in our Forums,
or
Contact Us and let us know.
Follow the Discussion
I want it in the worst way too.
For example, when you hover over a photo, the photo highlights and gets bigger... but how about the photos next to it? The single huge picture obscures all the pictures around it! Typically when I look at a single picture, I also want to look at its surrounding pictures. Obscuring the surrounding pictures make it difficult for me to determine whether I care about the next picture to highlight it too or just to skip it. The current implementation would force me to highlight the next picture just to determine whether I care about it or not.
The transitions are very cool, nice work! However, I also hope there is a way to disable transitions. If I want to do something quick, transitions only make the process longer and more tedious.
Overall I think the app is cool and has potential. However more UI studies would make it even better imo. I have similar complaints about Picasa too, so it's not that I don't like this app, I just like better designed apps.
It has revolutionised my photo collections.
It was the first Application and UI that has really made me sit up and go "Yes" in a long time. I just wish it would accept non visual media files such as word docs. I guess I will have to wait for Longhorn for that.
Better still there is a free version which from memory has all the functionality but some restrictions on the number of custom tag groups you can create.
Charlie
What are the chances of getting this in Longhorn?
If they're low would a petition help?
I think it would make a very usable digital photo organizer. With all of the great things I'm hearing about Longhorn, I was expecting user interfaces like this in Longhorn. I'm not sure whether or not I was wrong to expect them but I hope that 2 years from now Longhorn has these capabilities. It would be an entirely new meaning to "My Pictures," "My Videos," etc. Heck, with such capabilies you could just make one folder (for video/pictures) called "My Media." It would give people a huge reason to upgrade/switch to Longhorn.
The entire reason why you're hovering over it is because you want a closer look at it, not because you're going through one photo at a time. I think that the automatic enlargement is an excellent feature. For people that don't, Microsoft could have an option that controls how big the photo gets and perhaps an option to turn it off.
I really like the transitions and think that they make the interface flow nicely. When this is a product, I'd like to see this in it.
Personally I think that part of the problem you're having is that the presentation was poor. Check this one out:
http://news.com.com/1606-2-5234982.html
That said, I will confirm that 1+1 does appear to equal 2. Hint, hint. Wink, wink.
Oh well ....
My guess is that if they did ship it tomorrow, it might have (on the side of the cardboard box) "System Requirements: Windows Longhorn 2006 (or later)."
Still, it is a very cool app. I'd like to see these concepts extended to browsing items other than pictures and movies -- the next "Windows Explorer," if you will. A lot of those functions -- sorting, instantly filtering, zooming in on a selection -- apply to all documents, not just "media." (X1 3-D, anyone? Or should they call it "Direct X1?")
The movie thumbnails should actually play a short loop of the movie (a la DVD title menus).
Music thumbnails should show the album art and play the song on hover.
If MS can get to work with hand/arm gestures and/or data gloves, then I can pretend to be Tom Cruise in "Minority Report" when I'm searching for my expense reports. LOL
Does that mean that it will be in Longhorn from what you know at the moment but will not guarentee it?
I always thought that DirectX was a graphics API.
Longhorn is going to be a great desktop OS for organizing, managing, viewing, and sharing photos. You will see more scenarios like these in the demos in the coming years as Longhorn matures. Stay tuned
God i registered just to respond to this....
Wheres the glass screens to go with it i mean really technology needs to take a step up and microsoft normally was the one to make those steps.
This is THE most amazing video i have seen demonstrating what i hope to be the future of file management. As was stated previously images and video just to name a few... And Zoomed previews of the video would be so nice
But really i agree with one other thing, why isnt microsoft backing devolopment of finger tip navigation etc like in minority report come on guys we all know that with situations like the 3d stuff demoed we'd all just kill for a pair of finger tip / palm mounted wireless input devices!
We are doing research into finger/gesture navigation.
http://research.microsoft.com/displayArticle.aspx?id=625
http://research.microsoft.com/%7Eawilson/
To answer a couple of points made earlier, we aren't "sitting on" these technologies. MSR works with all of MS's product groups to move technologies and designs over into products. At the same time, they are the experts on building products, not us, and they ar emuch better at the spit-n-polish than we are. While we do user studies, they do full-fledged usability tests as part of their iterative design and development processes. They know how to test products, and they have all the infrastructure to do it well; we don't.
We have to respect their development processes. Sometimes we put out prototypes directly from MSR, but they will never have the professional quality that a MS product has.
It's a good partnership. It works. It just takes longer to do it right than to do it wrong. We prefer to do it right, because that makes our customers happier in the end.
MS continues to play vital role in taking the world towards Next Generation...
Karim, I agree with You on several points, especially on applying this thechnology to Windows Explorer and all files, not just media files.
How about, in the "My Favourites" folder, instead of a list of URLs we get miniature current captions of the actual webpage?
That would be amazing to have, thoght it was kinda, uhm, not public yet, but I guess I can say at least that I like it very very much
What about algorithms that finds out if a persons name is a female or male name (without some kind of name database)
Perhaps as a part of Longhorn.
I've also developed some image analysis and archiving apps but never anything so cutting-edge. I especially like the "bin" style grouping, which allows you to select and batch process collections of images. But annotating images using hand tagged metadata is still too much work. It doesn't scale from a user point of view when you have not hundreds but thousands of images, such as video image captures. That is where pattern recognition algorithms become a necessity.
What I would like to see is an interface for these types of detection algorithms. In the video there is a button for face detection, indoor/outdoor recognition etc. It would be nice to generalize that capability. Find every picture with a pattern taken from a sample image, a pattern that could be a face, a car, etc..
ok, hope to see more vids of the new media group!
aq
Yes I agree the reason why you're hovering over a photo is because you want a closer look, but do you ever want to just look at a single picture? Typically, even in this app, pictures are grouped together because they are related somehow. It could be that they are related by date, or because they are pictures of the same trip. How do you know if you want a closer look at the next picture if you don't even know what the next one is?
Look at the current My Pictures implementation in XP, nothing is ever obscured. Look at the Filmstrip view of the pictures, the design puts the larger image above the smaller one, making sure the user can easily determine whether they want a larger view of the next one. Granted the current "My Pictures" implementation is not as "cool" as this one, it does get the job done simply and easily. If you have ever seen a non-techie use a computer, that's ALL they care about. They don't care about the flashy graphics or the smooth transitions, they only care that they can look at their pictures simply and quickly. Do you want to try to explain to your grandfather how the pictures map to surfaces of cubes that get translated and rotated at a certain angle then finally sheared to provide this perspective view?
I like this implementation. I like the fancy transitions and the nice 3d effects. I hope the scripting engine stays so I can write scripts to provide my own fancy layouts and transitions. But then I'm a developer like most people on this site. However, I have met too many non-tech people out there to know that they just don't care about all the fany features. They want something simple, quick, and in the end gets the job done. Anything beyond these goals just make the job more difficult and detracts from the overall experience.
Isn't it though. I always love those types of answers.
My guess is that since it is probably going to come out with longhorn, and longhorn is the first OS to have the .NET Framework at "heart" of the OS. And, all the new API calls being put into Longhorn are supposedly going to be natively .NET. I think you guys can tell where I am going with this... I would guess that this app was written in C# utilizing DirectX 9.
Very cool, always wonted the chance to do stuff like this.
The time relevance (temporal) capabilities are amazing, I currently use Adobe Photoshop Album, but this is way cooler, the thing I really miss is location based (spatial) capabilities e.g. ability to tag images with location data, don’t mind doing this manually via an Autoroute or map tie in, tell all cameras have GPS receivers..
This would allow me to bring up a map of the world/country with push pins in it of locations that i visited, which i find more relevant for searching than date/time, as usually you show people pictures of a certain location or trip to a certain place, though it might have been taken a year ago..
Nb. I currently organize my photos into folders based on location e.g. US > California > San Francisco.
Kaz
Chris, you probably want to see the previous video about the VIBE group.
Imagine not only the picture you are hovering over is enlarged, but its close neighbors are also enlarged, just not as much. Other pictures in the sorroundings are squeezed to allow for the necesary space, or perhaps they just slide to the sides.
Sounds cool to me. All you need to do is find out what kind of patents Apple has on it
This is great! I can't wait to use it. I have a few suggestions, take them for what their worth:
1) I've seen quite a few image management programs built around a single feature. I've also seen people switch programs or use multiple programs because of missing features. It would be nice to have some sort of plug in architecture to allow others to extend what you have. For example, ophoto or shutterfly could provide a plugin to allow images to be uploaded to thier print servers. Someone else might create a filter that creates bins of 'textures' or black and white images. Another plugin might launch an image editor or perform simple filters on the image.
2) I can imagine my list of keywords wouldn't fit in a context menu. It might be nice to have keyword groupings. Say I had a keyword for each member of my family. They could be in a keyword group called 'family'. Then I could find pictures of the family, or just my brother. In addition this would keep me from having to search a huge scrolling list of keywords.
3) I'd like to see keywords applied by drag and drop rather than right-click. It may sound silly, but I imagine it would be faster (like sorting mail).
4) With all feature rich applications, it would be nice to be able to remove or hide features. MS has played with this by creating menus in Office that hide seldom used features in the menus until you pause for a few seconds (or click on the down arrow at the bottom of the menu). This doesn't seem like the best solution to me. There are a good deal of features that I know for sure I will never use (I would never email my images, for example). By turning off features, I think it would make the features I do use easier to find (it would certainly help my Mom).
Just suggestions. Take em or leave em as you see fit.
You don't HAVE to wait for anything.
Look at the videos closely. The theme on both of them is the default XP theme with the green start bar. All of the default themes in Longhorn (unless some really early version I've never used) have a completely different look than the default for XP. So the evidence suggests that this was ran on XP meaning the application can be ran on my system.
I have no proof of what language it was in though C# makes a lot of sense. DirectX was definitely used which is apparent by the visual effects. Avalon has this power but evidence doesn't suggest this was done in Longhorn though the technology really does "fit" in the Longhorn space.
Where am I going with this? You have videos you have concepts and even though you don't have exact algorithms you have the basics of the application. What is stopping Joe Schmo Hobbyist from creating their own version of this application NOW? Nothing. Absolutely nothing. Well, there's the whole time = energy = money and Microsoft Research has a lot more than Joe Schmo Hobbyist will ever have.
The things I consider basic and a necessity are as follows:
1) Categorization. This is a no-brainer and isn't present in XP to my knowledge in My Pictures
2) The use of Bins.
3) Mouse-over zoom. Unless you can specify the thumbnail size this is a necessity
4) 3D View. It's very cool and though it's incredibly hard for my mind it's not incredibly impossible in today's technology.
I would only add the ability for the app to upscale the thumbnail size as less images are displayed. It did this somewhat and I think there is a threshold but the ability to change it and have bigger images would be ideal. People don't always have 20/20 vision.
I know I may get some flack for posting this and that's perfectly acceptable. I just don't believe technology this helpful should be slated for a "sometime later, if later" release. I don't like reinventing the wheel at all but I would support an interim product that had limited capability I've described. Necessity is trumping my patience on this one and it may get me started on a simplistic version. This is also the type of programming challenge I've been looking for. It's extremely hard on the surface yet because it's so cool it would actually be fun to spend countless hours on.
I can see my About box now: "Concepts: Microsoft Research Next Media (link to homepage).
Implementation: Me. If you like this product, please support the Microsoft Research team by obtaining their version.
Microsoft: Please don't hurt me. By releasing this now, people are drooling over your version which will be better than mine could ever be. I just couldn't wait any longer, so please don't hold it against me"
Looking closely it almost appears as so, but I can't get a clear view at the taskbar or the title to see if it is in fact ran in a VPC session. (Nor have I ran VPC to know what it looks like)
Just because it "feels" Avalon doesn't mean it is and something could easily do this now using DirectX. I couldn't see how someone or a team of people could crank out something this good so close to Longhorn's initial alpha release. I guess it is possible it's just highly unlikely.
If it is done in Avalon it still shouldn't discourage anyone from producing their own version if they so choose. This can be done now without the need to wait until Longhorn or possibly after or possibly never, whichever comes first.
I would move the mouse if I wanted to see what was above the picture I was currently viewing.
It was said that they developed it with Avalon in mind so I would think that it could ported to Avalon. And if in its current implementation it can't, there is always another way of coding things to accomplish the same task. ^_^
There is a category view in Windows XP but it isn't this advanced. Right click>Arrange Icons By>Show in Groups
Please!
bonj
Here I thought that they had discontinued Picture It! Express. I still have Picture It! Express 2.0. It came with MS Works 99 suite.
http://www.wired.com/news/technology/0,1282,64353,00.html?tw=wn_tophead_3
Just curious if there were "other" areas of research that could be of interest (for videos).
so what about computers, "search for all items with a box" and it could interprit that in a complete unuseful manner. I would like to be able to select a section of a pic, then play with it, morph it, inverse it, highligh 'what i see' and then say "searh for simliar" - that would then let me search for what i see.
see what i mean?
samuel
Looks awesome.
On the image detection -- one thing I really find tedious is flipping pictures back to the right orientation when I've rotated the camera... be great if it included some detection algorithm to detect and auto-rotate..
I had a question. I don't quite understand between face detection and face recognition.
From what i understand from the video, you can extract pictures with faces in there which means face detection.
How about face recognition? You mean can extract pictures with identical faces?
Eeks, isn't that requires some processing but when he demos it, a bit too fast ... can anyone explain to me a bit more please? I love this technology
I've always wondered why cameras don't have some kind of mercury switch mechanism that determines the correct image orientation during acquisition.
I see no reason, though, why this would not be done in the background and stored as metadata with the image. An image would then have a metadata list of number of faces in the image, their location, and then each location would have some recognition structure (eigenface or whatever). If the face has been confirmed as someone in the contacts list, it there could be a pointer to a contact in WinFS.
Processing new images could be scheduled and done with unused cycles.
Samuel
The site for it is www.microsoft.com/imaging.
You can also watch a video that shows how they organize photos at
http://www.microsoft.com/products/imaging/docs/WatchVideo.html
We worked on a project not dissimilar than this for a film called Bodysong... it's basically a film made up of archive footage to show someones life...
We built a database that allowed you to fly through the different chapters using a bespoke 3d engine.
http://www.bodysong.com
This looks like the natural porgression of it... nice!
we also ran our database of images through a program made in director that would pick a key colour for each image... therefore allowing the stream of images to appear beautiful before the image for that section streamed in.
our company site can be seen at http://www.engagestudio.com
Microsoft Research seem to do an awful lot of great things - 'Singularity' is another example.
Remove this comment
Remove this thread
close