Archive for the ‘Videocams’ Category

Forrester ranks Polycom top for video collaboration

Wednesday, August 22nd, 2012

In a new independent report from Forrester Research, Polycom has been ranked top for video collaboration strategy and current offerings.

Cisco was ranked a close second while Huawei was out of the top three.

The Forrester Wave: Room-Based Videoconferencing, Q3 2012 Report

The Forrester Wave: Room-Based Videoconferencing, Q3 2012 Report

Polycom received perfect scores in the categories of quality of experience, interoperability, product roadmap, partnerships, security and compliance.

“Polycom has a diverse portfolio of endpoints including a range of options that run the gamut from mobile and room-based systems to immersive telepresence,” wrote Phillip Karcher, the author of the Forrester Research report.

Polycom is a global leader in open standards-based unified communications (UC), with a global ecosystem of over 7,000 partners.

The report, titled “The Forrester Wave: Room-Based Videoconferencing, Q3 2012”, uses 39 criteria to evaluate the top vendors in the industry, looking at video collaboration solutions for every work environment – from mobile and desktops to room systems and immersive telepresence – along with the services that support them.

The Forrester Wave report also discusses the growing interest in “face to face” video collaboration, which is helping businesses increase productivity, improve employee engagement and trust and lower costs.

“Polycom will also continue to offer the industry’s broadest multi-vendor integration and lowest TCO. For our more than 415,000 customers globally, Polycom is delivering innovations in video collaboration that are changing the way people work, collaborate, learn, govern, create and even help save lives,” said Andy Miller, CEO Polycom.

Polycom RealPresence video collaboration solutions deliver secure, high-quality video to companies of all sizes, so that teams can meet face-to-face with colleagues, partners and customers anywhere people work ― in the office, at home, on the go, or on the manufacturing floor, in a hospital or, at a movie set.

The RealPresence Platform provides the robust provisioning, management, scalability, reliability, flexibility and security required both for video enterprises and their private video clouds serving thousands of users, as well as for Video-as-a-Service (VaaS) delivered from the cloud potentially to tens of millions.

Read more about the Polycom RealPresence Video Collaboration Solutions below.

(more…)

Game-changing technology: wearable unified displays with scalable imagery

Sunday, July 8th, 2012

In a time when new technology and gadgets are launched everyday, and every one of their makers proclaim them to be game-changers, it is truly heartening to see products showcasing technology that has genuine potential to become a game changer in the near future.

Wearable video displays are an example. The Epson Moverio and Google Glass are examples of such gadgets.

Wearable unified displays can be used everywhere and for any purpose in the future, here an aircraft engineer has his hands free while referencing his technical manual through an Epson Moverio.

Wearable unified displays can be used everywhere and for any purpose in the future. Here an aircraft engineer can have his hands free (if he clips the controller to his overalls) while referencing his technical manual through an Epson Moverio.

In my humble opinion, there are two main constraints in consumer devices that represent frontiers in product developments for gadget makers.

Google Glass was showed off by co-founder Sergey Brin (not in this picture) at the Google I/O conference in June.

Google Glass was showed off by co-founder Sergey Brin (not in this picture) at the Google I/O conference in June.

These are display technology and device input.

What defines a smartphone and a tablet, laptop and a laptop, or even a television and even the cinema?

It is the size and form factor of the device.

Apart from the backend processor and performance, a large determinant of their category is simply their screen size.

If it’s a 3-5 inch device (with 3G connectivity), it’s a smartphone. 7-10 inch flat screen (with or without 3G), it’s a tablet (used to be netbook not too long ago).

Beyond that, you have the 13-17 inch notebook computer, and still beyond that – the desktops with even bigger screens.

Sony HMZ T1 Personal 3D Viewer

Sony HMZ T1 Personal 3D Viewer

Need we go into the numerous sizes for TVs, projectors and erm – cinemas?

A large determinant of device portability and usability – which often run counter to each other – is the size of the display.

Greater portability results in more squinting, whereas nice expansive viewing pleasure means bigger heavier screens.

A viewer/display device that allows the user to divorce display size from processing hardware will free the user from having to own multiple devices with different form factors.

Imagine the ability to make calls, read ebooks, play games, work on documents or edit images/video, watch the equivalent of large-screen movies on a single viewer/display.

Wanna look like an X-men? Try the Cyclops look.

Wanna look like an X-men? Try the Cyclops look.

No need to purchase separate bulky devices with their own different-sized displays.

You’d still need to have the little black-boxes containing the different processing hardware and device input.

But imagine sharing the same viewer/display amongst all these devices.

All of them “plug” into the same viewer/display.

There’s no more difference between a smartphone and a 3G-capable tablet computer.

Cameras and camcorders could also pair up wirelessly.

Although the device closest to bringing this to reality is the mounted-visor display/viewer, I’m sure there’ll be other physical means and ways to achieve this “unified display”.

Olympus has just announced its wearable display prototype - MEG4.0.

Olympus has just announced its wearable display prototype – MEG4.0.

Eventually, I can envisage (pun intended) displays embedded in contact lenses or even implanted into the cornea of the eyes.

In fact, patents have already been filed for such designs.

I haven’t seen any patents that bypass the eyes and inject vision signals directly into the brain though.

In a way, nature already unifies human vision through our eyes. We do not put on a different set of eyes to see different things.

The Olympus MEG 4.0 promises to connect to devices via Bluetooth.

The Olympus MEG 4.0 promises to connect to devices via Bluetooth.

Everything we see are through the same pair of eyes.

Why should we require a different screen to view the contents of different devices?

Now this single unified display may not be good news to device makers, although it surely should be the dream of consumers – even if we may not realise it yet.

Surely it’s better to be able to sell a single user multiple devices with different screen sizes if his/her needs forces him/her to do so.

Many users today own separate tablets and smartphones.

Sony HMZ T1 Personal 3D Viewer

Sony HMZ T1 Personal 3D Viewer

Apple, for instance, would rather sell you an iPhone AND an iPad, than only a single mobile device that “throws” its display into a single viewer with a “scalable” image size.

Google Glass with augmented reality.

Google Glass with augmented reality.

I’m sure it isn’t even news to the makers.

It’s a bit akin to traditional airlines offering budget flights.

Why introduce new budget services – even though it would benefit customers – that will change the game and threaten their erstwhile profitable business?

Fortunately, we have entrepreneurs who have broken the ranks to offer such choice to budget travellers.

A closer analogy would be the way netbooks had cannibalised the mainstream laptop market.

Sony HMZ T1 Personal 3D Viewer

Sony HMZ T1 Personal 3D Viewer

We have Asus to thank for “creating” the netbook niche to bring affordable and no-frills portable computing to the masses.

The Epson Moverio and Google Glass are not the first such device on the market, and in their current form are not mature enough for mass adoption by consumers yet – in terms of usability and feature-set, as well as content and consumer readiness.

But hopefully it precipitates more momentum to force makers to get on the bandwagon.

The other frontier

The other “physical constraint” to device form-factor is input interaction, a large part of which is text-input.

The size of screen again comes into the picture again if an on-screen keyboard is used.

And if a physical keyboard is used – be it a BlackBerry style keypad or a bigger portable keyboard – it translates into the overall form factor of the device.

Voice-input represents a means to free a device from a physical or on-screen keyboard, and even provide interactivity – through voice commands and voice search.

It’s also still early days and much development is needed before we can enjoy keyboardless devices

But as consumers and end-users, we can always dream on.

* Photos and illustrations used in this article belong to Epson, Google, Sony, and Olympus.

窥探一下未来的显示浏览器。如今已有一些头目。譬如:Epson Moverio BT-100, Google Glass, Sony HMZ T1 Personal 3D Viewer, Olympus MEG4.0。希望有朝一日,能够只凭一个显示机利用多种电子设备器材。

Seminar: Canon Cinema EOS System in TV Production by Alexander Buono from Saturday Night Live

Saturday, July 7th, 2012

Thanks to Canon’s invitation, I got to attend this interesting talk by the Director of Photography for Saturday Night Live Film Unit.

Alexander Buono shared his insights and experiences from his tight weekly schedule and workflow for shooting the longest running series in American TV.

* All photos taken with a handheld smartphone.

Alexander Buono from Saturday Night Live on TV production using Canon Cinema EOS.

Alexander Buono from Saturday Night Live on TV production using Canon Cinema EOS.

At the Lido Cineplex in Singapore, Buono gave some very down to earth and pragmatic tips and shared the setups and settings that he uses with his Canon equipment.

One of Alex Bruno's setups.

One of Alex Bruno’s setups.

Buono related how he began shooting for the popular Saturday Night Live (SNL) program in 1999 using film.

In 2009, he began using the Canon EOS 5D Mk II, before including the EOS 7D and Cinema EOS C300 in his digital video inventory.

Some interesting nuggets from his presentation and Q&A below.

To better understand and play with what the settings on the Canon EOS C300, check out the C300 Menu Simulator which emulates the menu system on the camera.

Alex Buono's rig

Alex Buono’s rig

If you’re into shooting video, here are some people to follow on Twitter.

The Hobbit was shot in 5k 3D video at 48fps, while James Cameron’s next installment of Avatar is being shot in 5k 3D video at 60fps.

These will likely look hyper-realistic, departing from the more surrealistic feel of film.

Buono recommends shooting at 24fps for the film look, even though videos for TV are typically shot at 30fps.

Another of Buono's setups

Another of Buono’s setups

Buono recommends keeping the shutter speed constant throughout a shoot. A good setting would be the reciprocal of twice the frame rate for the shoot.

So for a video shot at 24fps, a good shutter speed to use would be 1/48 second.

Overall, it was an interesting session for the attendees who packed Hall 5 of the cinema in which the workshop was held.

Alex Buono attended the School of Cinematic Arts under the University of Southern California, majoring in film production and still photography.

Alex Buono, Director of Photography for Saturday Night Live Film Unit.

Alex Buono, Director of Photography for Saturday Night Live Film Unit.

He spent his early post-college years as camera crew for studio films including Twister, Conspiracy Theory and Armageddon, learning from top ASC cinematographers and personal mentors like Don Burgess and John Schwartzman.

He soon became Director of Photography of the Saturday Night Live Film Unit.

PC Show 2012 promotions: Canon

Thursday, June 7th, 2012

Canon will be at the PC Show 2012 at Suntec Singapore at Level 4, Hall 404 and 403, Booth A2016.

Canon's PC Show 2012 promotions for printers, inks, DSLRs, camera and camcorders.

Canon's PC Show 2012 promotions for printers, inks, DSLRs, camera and camcorders.

Here are seven brochures for Canon’s PC Show 2012 promotions for printers, inks, DSLRs, camera and camcorders.

PC Show 2012 promotions listing

Thursday, June 7th, 2012

This year’s PC Show 2012 is at Suntec Singapore from 7 June (Thursday) till 10 June (Sunday). Here’s a listing of some of the promotions and flyers for the four-day IT exhibition.

PC Show 2012 from 7 - 10 June at Suntec Singapore from 12 - 9 pm. Admission FREE.
Click on the links below to check out the deals and promotions from the brand. To see a consolidated list of postings related to PC Show 2012, click here. You can also click on the button near the top of the sidebar on the right that looks like the logo above.

Floor plans for Levels 4 and 6 can be found here.

New HD camcorders: Canon LEGRIA HF M56, M52, R38, and R36

Wednesday, March 28th, 2012

Canon has introduced its new LEGRIA camcorder lineup with social sharing capabilities. The new LEGRIA HF M56 (S$1,349), HF M52 (S$1,549), HF R38 (S$1,099), and HF R36 ( S$899) camcorders feature enhanced low-light performance and wireless connectivity.

Available in Singapore mid-April.

Canon takes LEGRIA social with its slew of 4 new camcorders.

Canon takes LEGRIA social with its slew of 4 new camcorders.

Canon Singapore today unveiled its new high definition LEGRIA HF-M and HF-R series that allow users to directly share their videos on social networking sites, without the need to transfer them to a PC.

Bring the LEGRIA HF-M camcorder underwater with the WP-V4 Waterproof Case.

Bring the LEGRIA HF-M camcorder underwater with the WP-V4 Waterproof Case.

Offering Full HD recording in a compact and lightweight body, the new camcorders offer consumers new features like Wi-Fi connectivity, the option to capture videos in MP4 or AVCHD file format, and full Digital Living Network Alliance (DLNA) certification to support the sharing of videos through DNLA-certified HDTVs, and mobile devices.

The HF-M series also features an enhanced version of Canon’s HD CMOS PRO sensor to deliver uncompromised professional-quality full HD video, even in extreme low-light conditions.

The Canon HD video lens and a powerful 51x advanced zoom in the HF-R series ensure that users can easily capture the action anywhere, even from far distances.

More detailed descriptions of the features can be found below.

(more…)

Panasonic camcorders: Spring/Summer 2012 line-up

Friday, March 16th, 2012

For camcorders, Panasonic has a line-up of 9 new models that will hit the shelves in April. These comprise the HC-X900 & HC-X900M, HC-V700, HC-V500 & HC-V500M, HC-V100, HC-V10, HX-WA20, and HX-DC2.

Panasonic's Summer/Spring line-up of camcorder ranges from the compact HC-V10 to the flagship HC-X900M.

Panasonic's Summer/Spring line-up of camcorder ranges from the compact HC-V10 to the flagship HC-X900M.

Panasonic’s nine new camcorders for Spring/Summer 2012 features advanced lenses; high sensitivity sensors; sophisticated Optical Image Stabilizer (O.I.S.) capabilities; and powerful image-enhancing engines.

HD Writer video editing software is included with all models.

Flagship HC-X900 and HC-X900M

Panasonic HC-X900

Panasonic HC-X900

The re-engineered F1.5 LEICA Dicomar lens of the flagship HC-X900 and HC-X900M enables bright image rendering, high corner-to-corner resolution and Nano Surface Coating Technology to significantly reduce ghosting and flaring, which minimises distortion and degradation of contrast.

The advanced 3MOS Sensor is capable of high-speed processing of massive number of pixels – approximately equivalent to four times the pixel count of Full-HD – using pixel shift technology with submicron accuracy, to produce high-detailed images with excellent colour reproduction.

Panasonic HC-X900M

Panasonic HC-X900M

The newly-developed, ultra-high speed Crystal Engine Pro II can also dramatically reduce noise by 40% compared to the previous models to record clear, stunningly beautiful images even in dim lighting.

The new Hybrid O.I.S.+ corrects horizontal, vertical, yawing and pitching movements made by the hand during recording; while its five-axis correction capability takes care of any rolling motions resulting from walking.

3D shooting is enabled by the 3D Conversion Lens, creating the realistic depth, texture and dimensionality of 3D.

HC-V700

Panasonic HC-V700

Panasonic HC-V700

The HC-V700 features a 28mm wide angle for recording in tight spaces such as indoors.

The 46x Intelligent Zoom extends ultra-telephoto shooting beyond the range of the 21x optical zoom; while the HYBRID O.I.S.+ suppresses blurring across an entire zoom range even for handheld zoom shots in high magnification, to ensure crisp and beautiful images..

Images taken in 2D can also be converted to 3D with the new 2D to 3D Conversion Function.

HC-V500 and HC-V500M

Intelligent Zoom is up to 50x, using the LSI “Crystal Engine PRO” image processing engine.

Panasonic HC-V500 & HC-V500M

Panasonic HC-V500 & HC-V500M

Hybrid O.I.S.+ and five-axis correction are also available.

Advanced Highlight Playback function enables quick viewing of recorded images. The Intelligent Index System detects zooming, panning, scene changes and faces in recorded images as “highlights” and then automatically plays back the detected highlight scenes – and scenes that the user has personally selected – according to a set time interval.

HC-V100 and HC-V10

Panasonic HC-V100

Panasonic HC-V100

Both models are compact and lightweight, and feature high optical zooms stabilised by the POWER O.I.S.

The HC-V100 has an 42x Intelligent Zoom, as well as Power O.I.S., with a continuous recording time of up to 165 minutes on a fully charged battery; while the HC-V10 has an even more powerful 63x optical zoom lens that can be “enhanced” to 70x.

HX-DC2 (monitor folded in) & HX-WA20

HX-DC2 (monitor folded in) & HX-WA20

HX-WA20 and HX-DC2

For active outdoor use, check out the new waterproof (up to 3m) and dustproof HX-WA20.

It can record fast moving sports motion at 240 fps (640×360 px) or 480 fps (320×180 px) for 15 seconds, and then playback in slow-motion at 60 fps at ¼ or ⅛ of the original speed, allowing analysis of golf swings etc.

Both the HX-WA20 (16.4 MP) and HX-DC2 (14.4) shoot Full-HD, HDR stills and 5x optical zoom.

IT Show 2012 promotions: Sony

Friday, March 9th, 2012

Sony will be at the IT Show 2012 in Suntec Singapore almost every which way you look and on nearly every level of Suntec.

Check out Sony's promotions and deals at the IT Show 2011.

Check out Sony's promotions and deals at the IT Show 2011.

Lookout for deals and promotions for their Bravia TVs, digital photo frames, Home Entertainment Systems, personal music products like the Walkman, video and still cameras, tablets and Vaio notebooks and accessories such as power supplies, batteries and memory cards.

Digital photo frames and compact cameras

Digital photo frames and compact cameras

Videocams, DSLRs and MIL cameras

Videocams, DSLRs and MIL cameras

Tablets and Vaio notebooks

Tablets and Vaio notebooks

Personal music products

Personal music products

Bravia TVs and Home Theatre Systems

Bravia TVs and Home Theatre Systems

Power supply and memory cards

Power supply and memory cards

Click on the thumbnails above to view or download Sony’s promotion flyers for their different categories of products.

CES 2012 roundup: Sony gadgets for 2012

Saturday, January 14th, 2012

Here is a round-up of Sony’s electronic gadgets for 2012, as released during the Consumer Electronics Show (CES) 2012 in Nevada, Las Vegas, USA.

Sony's line-up for 2012 as announced during CES 2012.

Sony's line-up for 2012 as announced during CES 2012.

The CES is a chance for consumers to get a glimpse of the gadgets lining up to hit the market in the near to mid-term future.

Sony’s line-up for 2012 includes the new Handycam range; the Cyber-shot W-Series and S-Series; the NWZ-Z1000 series (Sony’s first ever Walkman powered by Android); and a new line-up of BRAVIA  televisions.

Here’s a list of these new products.

Handycam

Sony HDR-PJ760VE Full HD Flash Memory camcorder

Sony HDR-PJ760VE Full HD Flash Memory camcorder
  1. HDR-TD20VE Double Full HD 3D Flash Memory camcorder
  2. HDR-PJ760VE Full HD Flash Memory camcorder

New camcorder accessories

  • RDP-CA1 Portable Speaker
  • HVL-LE1 Battery Video Light
  • VCT-MP1 ‘Multipod’
  • 22MB/s SDHC Class 10 media cards
  • World’s first Balanced Optical SteadyShot for even more stable images while walking/zooming (HDR-PJ760VE)
  • Increased built-in projector models led by the HDR-PJ760VE capable of displaying 100-inch screen size (diagonal)
  • The world’s smallest and lightest 3D Handycam® (HDR-TD20VE)
  • Clearer dialogue with Auto Wind Noise Reduction and Closer Voice (both models)
  • By Pixel Super Resolution processing for high-resolution still images and Extended Zoom (both models

Cyber-shot

Cyber-shot compact digital cameras

Cyber-shot compact digital cameras
1. W-Series

  • W670
  • W650
  • W630
  • W620
  • W610

2. S-Series

  • S5000
  • All-round images with new 360 Sweep Panorama
  • SteadyShot image stabilisation for clearer handheld shots
  • Intelligent Auto mode recognises up to 33 scene settings for great results automatically
  • Picture Effect modes for easy in-camera effects

Walkman

Sony Walkman NWZ-Z1000 series

Sony Walkman NWZ-Z1000 series
NWZ-Z1000 series
  • First ever Walkman® powered by Android™
  • Unique ergonomic body design
  • Features the new S-Master MX digital amplifier which greatly reduces noise and distortion level in your music
  • A new distinctive W.BUTTON allows one to launch the unique W.Control app; allowing one to explore their music collection in an intuitive manner
  • Available in 8GB, 16GB and 32GB capacities

BRAVIA

Sony BRAVIA TV

SonyBRAVIA EX650-series LED LCD HDTV
  1. HX855-series LED LCD HDTV with 3D
  2. HX750-series LED LCD HDTV with 3D
  3. EX650-series LED LCD HDTV
  • Two distinct series –the flagship HX and the mid-level EX
  • Brilliant Full HD (1080p) picture quality with Dynamic Edge LED backlight
  • Picture qualities further enhanced with Sony’s X-Reality PRO and X-Reality digital video processors
  • They are Skype™ embedded, where users can enjoy free widescreen Skype-to-Skype video calls. Users can also make voice calls while simultaneously watching TV.
  • Sony’s HX855 carries forward Sony’s distinctive Monolithic design incorporating Gorilla® Glass by Corning®, allowing for thinner, lighter and stronger screen material.
  • Complete with X-Reality PRO and Motionflow XR 960, the HX850 is a fully Internet-connected television with built-in 3D, delivering a premium viewing experience when on, and a stunning addition to home décor even when off.

Behind-the-scenes story about Microsoft Kinect for Xbox 360

Saturday, July 9th, 2011

Dr Jamie Shotton had joined the Machine Learning & Perception group at Microsoft Research Cambridge (MSRC) in June 2008 as a post-doc for a few months when he was roped in by the Xbox product group to help launch the product by Christmas 2010.

He shared the experience with 4th year undergraduate Engineering students at the University of Cambridge Engineering Department earlier this year.

The body was divided into 31 different body parts to be recognised and reconstituted into a human pose.

The body was divided into 31 different body parts to be recognised and reconstituted into a human pose.

I was browsing through the university’s newsletter last week when I came upon this interesting story about some of the developmental challenges of the Microsoft Kinect for Xbox 360 and how they were surmounted. You can read the full original article here. Images used in this posting are from the original article.

The Kinect for Xbox 360 is a motion sensing input device for the Xbox 360 game console. Based around a webcam-style add-on accessory for the Xbox 360 console, it allows users to control and interact with the Xbox 360 without the need to touch or hold a game controller such as a joystick – depending instead on bodily gestures and spoken commands.

Dr Jamie Shotton from the Cambridge research laboratory in the UK

Dr Jamie Shotton from the Cambridge research laboratory in the UK.

Shotton now works for Microsoft at their Cambridge research laboratory in the UK. He had completed his PhD research in computer vision from 2003 to 2007. His initial research at the MSRC was on automatic visual object recognition – teaching computers how to recognise different types of objects in photographs such as cars, sheep and trees.

“Little did I know at that point how quickly I would get pulled into the frenzy of research and development around Kinect, and how this blue-skies research could be applied to such a practical problem,” Shotton recalled.

Enabling tools

At the point that Shotton was invited, Microsoft had already developed a few enabling tools.

Shotton's research into automatic visual object recognition trained computers to recognise different objects in photographs.
Shotton's research into automatic visual object recognition trained computers to recognise different objects in photographs.
Shotton's research into automatic visual object recognition trained computers to recognise different objects in photographs.
Shotton's research into automatic visual object recognition trained computers to recognise different objects in photographs.

Shotton's research into automatic visual object recognition trained computers to recognise different objects in photographs.

Depth-sensing camera. The new Kinect camera worked at 320×240 pixels and 30 frames per second versus other depth cameras at very low resolutions of 10×10 pixels.  “You could even make out the nose and eyes on your face,” “Shotton observed. The better depth accuracy helped with human pose estimation by eliminating objects in the background since they were further away. The colour and texture of clothing, skin and hair could also be normalised away. The depth camera was “active”, illuminating the subject with its own structured dot pattern of infra-red light so that the camera worked even in the dark.

Prototype human tracking algorithm.  The algorithm constantly compares its predictions of the body’s movements with the actual movements and then makes adjustments to improve the accuracy of its predictions.

Showstoppers

The tracking algorithm suffered from three limitations. First, the subject had to stand in a T-pose for the algorithm to lock it in initially. Second, if the subject moved too erratically and therefore unpredictably, the algorithm would lose track and would not be able to recover until the subject returned to the T-pose for recalibration. This could happen as often as every 5-10 seconds. Finally, the algorithm only worked with the limited number of body sizes and shapes that it had been trained with. Shotton’s mission was to overcome these showstoppers.

Overcoming the limitations

To allow the algorithm to recognise a subject and its posture without having to start from a T-pose, Shotton leveraged a fellow researcher’s (Dr Stenger) technique called “chamfer matching”: the subject’s image was compared with a training database of body images and once the closest match was selected, the 3D data for that match could then be utilised as the human pose for the subject.

However, there was an astronomical number of human poses based on the different combinations of position and orientation of body parts such as the arms, legs, knees and ankles. Shotton divided up the body into 31 parts so that each of the parts could be matched independently before building up the skeleton and body pose from the position of these parts. This was where Shotton’s PhD work on object recognition came in handy.

Although this substantially reduced the size of the image database needed to train the algorithm, the training database was still huge. The team had recorded hours of footage at a motion capture studio with several actors doing “gaming” moves such as dancing, running, fighting and driving.

The millions of training images would have taken months to train the algorithm. The team got help from colleagues at Microsoft Research in Silicon Valley who had developed an engine called “Dryad” for efficient and reliable distributed computation. Using a cluster of 100 powerful computers, the training time was reduced to less than a day.

Read the details of Shotton’s experience in the full original article here.