Saturday, December 29, 2012

Pthreads Win32

Pthread is the POSIX 1003.1-2001 standard which defines an application programming interface (API) for writing multithreaded applications. It is commonly used in linux. Win32 does not and is unlikely to ever support pthreads natively.

A good comparison between threading mechanism in Win32 and POSIX systems is Robert Sayegh's presentation "PThreads Vs Win32 Threads". Clay Breshears believed that Pthreads are better than Win32 threads

The Pthreads Win32 project seeks to provide a freely available and high-quality solution to this problem. The current release version is 2.9.1 (2012-05-27). It can be compiled with Visual C++Some tests suggested Pthreads Win32 is faster than Win32 threads.

The source tree and precompiled .DLL, .LIB and necessary header files are included in the zip file named "pthread-w32-v-v-v-release.zip" at:

ftp://sourceware.org/pub/pthreads-win32 


(v-v-v is the version number, e.g. 1-4-0, and “release” is a descriptive term that may vary, e.g. beta1, rc1. Etc). Versions older than 2.9.1 are in self-extracting .exe files.


Just the DLLs, LIBs, header files, and admin documentation is available at:

ftp://sourceware.org/pub/pthreads-win32/dll-latest


The source tree is available as a gzipped tar file from:

ftp://sourceware.org/pub/pthreads-win32


The source files are also available in unpacked form at:

ftp://sourceware.org/pub/pthreads-win32/sources

Wednesday, December 19, 2012

TI Shifting Focus to Industrial and Automotive Area

Texas Instruments announced last month that it is eliminating 1,700 jobs worldwide as it cuts spending in its wireless business. As the company's biggest smartphone and tablet customers have begun developing their own chips, Texas Instruments is shifting the focus to industrial and automotive customers. Yesterday it announced that it will lay off more than 500 people at a research and development plant near Nice, France, in the coming months. The plant was primarily dedicated to developing OPAM microprocessors used in smartphones and tablets.

TI's automotive products include Automotive Central Body ControllerDriver Information / Entertainment which features Active Noise Cancellation (ANC), instrument cluster, infotainment as well as digital radio solutions.

In July 2012, CSR PLC sold its handset connectivity and location operations and the associated technology to Samsung and shifted to voice & music, automotive infotainment, indoors location finding, imaging and Bluetooth smart.

Tuesday, December 18, 2012

Allegro DVT FPGA based HEVC (H.265) Decoder

At IBC 2012, Allegro DVT will demonstrate an FPGA based, real-time, high definition HEVC (H.265) decoder, according to Allegro DVT.

Monday, December 17, 2012

Top 10 Semiconductor Vendors by Revenue, Worldwide, 2012


Table 1. Top 10 Semiconductor Vendors by Revenue, Worldwide, 2012 (Millions of Dollars)
Rank 2011
Rank 2012
Vendor
2011 Revenue
2012 Estimated Revenue
2011-2012 Growth (%)
2012 Market Share (%)
1
1
Intel
50,669
49,295
-2.7
16.6
2
2
SamsungElectronics
27,366
24,974
-8.7
8.4
6
3
Qualcomm
9,998
12,954
29.6
4.4
4
4
Texas Instruments
11,754
11,001
-6.4
3.7
3
5
Toshiba
11,769
10,162
-13.7
3.4
5
6
Renesas Electronics
10,650
10,030
-5.8
3.4
7
7
STMicroelectronics
9,635
8,410
-12.7
2.8
8
8
SK Hynix
9,388
8,340
-11.2
2.8
10
9
Broadcom
7,160
7,792
8.8
2.6
9
10
Micron
7,643
6,935
-9.3
2.3

Others
150,811
147,657
-2.1%

49.6


Total
306,843
297,550
-3.0
100

The mobile phone chipmaker saw its revenue rise 29.6% year over year in 2012, while the overall market declined 3%. Of the top 10 chip vendors worldwide, only Qualcomm and No. 9 Broadcom increased their revenue this year. Both Qualcomm and Broadcom are Apple iPhone suppliers. 

Qualcomm surged to become the No. 3 semiconductor vendor worldwide this year by revenue. Qualcomm bucked the industry trend and grew this year thanks to continued adoption of smartphones and the growth of 3G and LTE technology in emerging regions, such as China and India, Gartner said

Broadcom's revenue rose 8.8%. While part of the company's year-over-year performance gain was attributed to the acquisition of NetLogic Microsystems, the highest organic growth in the company was once again achieved by the Mobile and Wireless division, which grew close to double digits. Broadcom's third division, Broadband, bounced back from a revenue decline in 2011 by growing in the mid single digits in 2012. 

Chip sales are expected to bounce back in 2013. Gartner forecasts global chip revenue to rise 4.5% to $311.4 billion next years, as IBD reported.

Sunday, December 2, 2012

H.265 (HEVC) and FPGA

High Efficiency Video Coding (HEVC) or H.265 is the current joint video coding standardization project of ITU-T Video Coding Experts Group (ITU-T Q.6/SG 16) and ISO/IEC Moving Picture Experts Group (ISO/IEC JTC 1/SC 29/WG 11). The current version is High Efficiency Video Coding (HEVC) text specification draft 8HEVC software repository is mainly at Fraunhofer Heinrich Hertz Institute.

It is expected H.265 can obtain the same quality as H.264 at half the bit rate. Cisco published a demo at Youtube to show the H.265 capability of high quality video with great bandwidth. Moreover, H,265 can cover both scenarios and support a wide range of applications, from DTH transmission to acquisition and storage.



According to Cisco's latest Global Mobile Data Traffic Forecast, world mobile data traffic to explode by factor of 26 by 2015, and two-thirds of the world's mobile data traffic will be video by 2015.


Many initiatives are underway for greater video quality. Japan's NHK is working on Super Hi-Vision video format , a successor to the familiar high-definition broadcast system, using 4K × 2K TV with approximately four times the pixel count of Full HD. Fujitsu announced that it has developed h.264 based codec equipment for systems transmitting Super Hi-Vision (SHV) video, with resolution 16 times higher than Japan's current Full Hi-Vision video. As entering an era of video data exploding, 3D and other high-resolution imaging, video compression technology is going to be more and more important. H.265 is expected to meet this trend need.

The technical architecture of HEVC is basically conceptually similar to H.264 (and 
prior standards):
  • Block-based
  • Variable block sizes
  • Block motion compensation
  • Fractional-pel motion vectors
  • Spatial intra prediction
  • Spatial transform of residual difference
  • Integer-based transform designs
  • Arithmetic or VLC-based entropy coding
  • In-loop filtering to form final decoded picture
Lots of variations at the individual “tool” level include:
  • Coding unit tree structure (8x8 up to 64x64 luma samples)
  • Prediction units (N=4, 8, 16, 32, shapes: 2Nx2N, NxN for smallest; for inter also 2NxN & Nx2N)
  • Transforms can cross prediction unit boundaries for Intra; not for Inter
  • Transform unit tree structure (maximum of 3 levels) Transform unit tree structure (maximum of 2 levels)
  • Transform block size of 4x4 to 32x32 samples (always square)
  • Angular intra prediction (17 directions for 4x4, 3 directions for 64x64, 34 directions for others)
  • Luma motion compensation: 1/4 sample precision, 8x8 separable with 6 bit tap values
  • Chroma motion compensation: 1/8 sample precision, 4x4 separable with 6 bit tap values
  • Advanced motion vector prediction with motion vector “competition” and “merging”
The following table shows the H.265 tool set and compares with H.264:

AVC High profileHEVC High efficiencyHEVC Low complexity
16 × 16 macroblockCoding unit quadtree structure (64 × 64 down to 8 × 8)
Partitions down to 4 × 4Prediction units (64 × 64 down to 4 × 4, square intra/inter + non-square inter)
8 × 8 and 4 × 4 transformsTransforms unites (32 × 32, 16 × 16, 8 × 8, 4 × 4 intra/inter + non-square inter)
Intra prediction (9 directions)Intra prediction (17 directions for 4 × 4, 3 directions for 16 × 16, 34 directions for rest)
Inter prediction luma 6-tap + 2-tap to 1/4 pelInter prediction luma 8-tap to 1/4 pel
Inter prediction chroma bi-linear interpolationInter prediction chroma 4-tap to 1/8 pel
Motion vector predictionAdvanced motion vector prediction (spatial + temporal)
CABAC or CAVLCCABAC (Context Adaptive Binary Arithmetic Docing)CAVLC (Context Adaptive Variabl Length Coding)
8b/sample storage and output10b/sample storage and output8b/sample storage and output
Deblocking filterDeblocking filter
-Adaptive Loop Filter (AFL) and Sample Adaptive Offset (SAO) filterSample Adaptive Offset (SAO) filter
Table 1. The toolsets provided to implement HEVC encoders can be used and modified to improve bit rate, video quality or both. Information courtesy Matthew Goldman, Ericsson, from the paper “High Efficiency Video Coding (HEVC) - The Next Generation Compression Technology,” presented at the SMPTE 2011 Technical Conference and Exhibition, Oct. 25-27, 2011.
Since technologies move so quickly, it makes it nearly impossible for broadcast engineers and designers working to get ahead of the competition to rely on application specific standard products (ASSPs) as the backbone of their hardware solutions. For this reason, Xilinx  ROBERT GREEN AND AARON BEHMAN believed that FPGAs offer the only viable platform for the next several years for companies hoping to exploit advantages of HEVC, since FPGAs can support the HEVC standard in software with hardware acceleration blocks for motion estimation and CABAC/CAVLC, which enables tradeoffs in device resource and performance while promoting design productivity. They said, "Depending on the application, the ability to trade off computational complexity, compression rate, robustness of errors and processing delay time are all elements that can only be evaluated in real-time with an FPGA-based design."

Tesla's 17" TouchScreen: The car dashboard of the future

Tesla Model S was named Motor Trend's 2013 Car of the Year, one of the automotive industry's most coveted awards.

Tesla Model S - Motor Trend Car of the Year

Tesla was co-found by Elon Musk who also is a co-founder of Paypal. He said, "Our aspiration with the Model S was to show that an electric car truly can be better than any gasoline car, which is a critical step towards the widespread adoption of sustainable transport. Nothing illustrates this more clearly than winning Motor Trend's Car of the Year by unanimous decision against a field of exceptional competitor." The field competitors included Porsche Boxster, BMW 3-series, Lexus GS, and Subaru BRZ.

It is very impressive that Tesla Model S will come with its 17-inch touch screen--a spectacular standard feature which, along with a smaller LCD display in front of the driver, replaces nearly all of the traditional controls and indicators short of basics such as the steering wheel and turn signals.

The Tesla S's touch-screen interface.
The Tesla S's touch-screen interface.
(Credit: James Martin/CNET)

According to Harry McCracken, "Tesla's screen resembles a large iPad or Android tablet--it's hard to overstate how overwhelmingly ginormous it looks--and sits in between the driver and front passenger. It has a music player, a navigation system, hand-free calling, a rear-view camera, climate controls, and other features which you might expect. It also lets you check up on the status of your car's battery, and--as you can see in the photo above--it sports a full-blown Web browser. And Tesla says that it will allow third-party apps, too."  It uses Nvidia's ARM-based Tegra platform running with Linux.

Tesla's mapping and music apps.
Tesla's mapping and music apps.
(Credit: James Martin/CNET)

Saturday, November 24, 2012

Nokia's Damian Dinning and PureView for the Lumia 920

Camera quality and features have been promoted by Nokia as key differentiation for its Windows Phone 8 handsets vs. competition from Apple (AAPL) and Android hardware OEMs.

Nokia's imaging chief Damian Dinning will leave the company on Nov. 30, 2012. He has been in charge of image and photography since 2004, and instrumental in the development of Windows Phone imaging software. Dinning's background includes roles at Minolta and Eastman Kodak, and Nikon.

Nokia's efforts to distinguish its smartphones with advanced photographic capabilities, introducing the PureView 808 with a 41MP rear-facing camera was Damian Dinning's PureView phase I development. His phase II was to improve low-light photography and eliminating camera shake for PureView for the Lumia 920.

Nokias Damian Dinning goes indepth on phase two of PureView for the Lumia 920 video

With this technology, its pictures were limited to 8.7-megapixels, and rather than pixel oversampling was used. Custom optics was developed to increase the aperture size to f/2.0 which was paired with a backside-illuminated CMOS and a high pulse power LED. The above image is taken with the PureView for the Lumia 920,  while below, there's a comparison shot taken by a Samsung Galaxy S III.

Nokias Damian Dinning goes indepth on phase two of PureView for the Lumia 920 video


The comparison shows the 920 does take much brighter, better images in low-light situations.

According to Daniel Cooper, the company customized Optical Image Stabilization (OIS) to reduce camera shaking. The OIS detects a camera's movement with a built-in gyroscope, compensating accordingly. Te entire optical assembly is moved mechanically to maintain parity with what you're shooting, which it claims is 50 percent more effective in reducing shake.

The following video compares OIS and non-OIS shot.



Wednesday, November 14, 2012

Omote 3D Printing Photo Booth

The world's first 3D-printer photo booth will be open on November 24th for business. Omote 3D, a pop-up store in Tokyo’s Harajuku district, is now creating portraits of customers via a 3D printer after scanning by the store.


Then it will transform the likenesses of persons into tiny, anatomically accurate, full-length figurines. 



Omote 3D offers discounts on parties of three or more, listed in the following table, but it’s still relatively expensive. For example, a single 10cm figurine of one person will set you back 21,000 yen, or $262 USD.

Table 1 The price of 3D Printed Figure

Solo2 persons3 people or more 
(per person)
(up to 10 cm)¥ 21,000¥ 42,000+ ¥ 16,000 people ×
(up to 15 cm)¥ 32,000¥ 64,000+ ¥ 21,000 people ×
(up to 20cm)¥ 42,000¥ 84,000+ ¥ 32,000 people ×



3d printed model of yourself in Japan

Monday, November 5, 2012

Q3 2012 Smartphone OS Market Share and Growth

According to new figures released from analyst firm IDC, Android shipments reached 136 million units in Q3 2012, which accounts for 75% of the 181.1 million shipments during the quarter. The 91.5% year-over-year growth was nearly double the overall market growth rate of 46.4%. 

iOS came in second place, but was the only other mobile operating system to claim double-digit market share in the quarter with 14.9% and 57.3% year-over-year growth rate. Note, however, that the iPhone 5 was released late in the quarter, and the full impact of its sales aren’t felt yet. But without a splashy new OS-driven feature like Siri in 2011 and FaceTime in 2010, the iPhone 5 relied on its larger, but not wider, screen and LTE connectivity to drive growth.

It is interesting that Microsoft Windows Mobile operating system surged 140 percent from last year, but only holds 2 percent of the total market share. With the launch of Windows Phone 8, it may performance better.

Other mobile OS, such as BlackBerry, Symbian, Linux, all are losting ground. 



Total Android units shipped totaled 0.7 million in 2008, but has grown to 333.6 million units year-to-date.

Sunday, September 23, 2012

Ambarella S2 IP Camera for 4K or 1080p120 H.264 Video

Ambarella enables video cameras with crazy 4K video resolution

Ambarella recently announced  S2, a security IP camera System-On-Chip (SoC) with support for the new 4K Ultra HD video standard. S2 is based on a Dual Core ARM® Cortex™-A9 CPU and a high-performance video engine that delivers multi-stream H.264 encoding up to 4K (3840x2160p30) resolution for outstanding image clarity and detail.  Hardware analytics acceleration combined with the dual-core CPU provide the processing power necessary to enable cameras to analyze video content and make intelligent decisions based upon it.  Leveraging low-power 32nm process technology and Ambarella’s proprietary system architecture, the S2 delivers 1080p60 encoding in less than one Watt and 4K video encoding in less than two Watts.

The S2 family provides an array of IP camera design choices.  The family is comprised of software and pin-compatible SoCs that range in performance from 1080p45 encode with a 500 MHz dual-core CPU to 4K Ultra HD encode with a 1 GHz dual-core CPU. Camera designers can also access advanced features including analytics hardware acceleration, panoramic lens de-warping, hardware face detection, digital Pan Tilt Zoom (PTZ) and electronic image stabilization (EIS). The Linux-based S2 Software Development Kit (SDK) supports a wide range of image sensors and enables features such as WiFi connectivity, cloud services, local SD Card recording and transcoding. The SDK also offers comprehensive 3A and image tuning tools for customization.


S2 Features
  • Image Sensor Pipeline with high-speed SLVDS/MIPI®/HiSPi™ interfaces, up to 32 Megapixel (MP) sensor resolution.
  • Wide Dynamic Range (WDR) with multi-exposure fusion to improve video in high-contrast lighting conditions.
  • Advanced 3D noise reduction with de-ghosting to enhance low-light performance.
  • 4K video (3840x2160p30 or 1920x1080p120) H.264 encoding to support the new Ultra HD standard. Multi-streaming to enable independent resolutions, frame rate and quality. Support is also offered for video transcoding and low-latency encoding. Main/High Profile with B-frames for high efficiency.
  • Dual Core ARM Cortex-A9 CPU with NEON and a 512 KB L2 cache provides headroom for customers’ applications, as well as enhanced power efficiency and real-time performance compared to single-core architectures.
  • Analytics hardware acceleration for efficient processing of tasks like object tracking, trip zone perimeter control and intelligent motion detection.
  • Real-time hardware-accelerated de-warping for 180/360 degree panoramic lenses.
  • AES/3DES/SHA-1/MD5 encryption engines.
  • Rich set of interfaces allows for compact designs and includes 32-bit DDR3L, Gigabit Ethernet, USBG 2.0 host and device, HDMI®, SDXC SD™ Card, and dual sensor interfaces.

Saturday, September 22, 2012

iOS 6, Xcode 4.5 and iPhone 5

According  to Anandtech.com,  ‘armv7s’, a custom Apple instruction set that adds VFPv4 support, a.k.a. vector floating point support, was added to Xcode 4.5 to support iPhone 5 which uses Apple’s new A6 SoC. Xcode 4.5 makes two major changes: it drops support for the ARMv6 ISA (used by the ARM11 core in the iPhone 2G and iPhone 3G), keeps support for ARMv7 (used by modern ARM cores) and it adds support for a new architecture target designed to support the new A6 SoC: armv7s.



Coronalabs.com made following suggestions. 

If you open up your Xcode project in Xcode 4.5, you’ll see something like the following. Note the presence of both ‘armv7′ and ‘armv7s’:

The main change you need to make is to make sure the builds for device only build armv7 (i.e. remove ‘armv7s’). Here’s what they should look like after:
In this setup, you are basically telling Xcode, that by default you still want the standard instruction sets, but for iOS, you are adding override behavior, forcing it to only build for armv7. Once we add armv7s support, you can easily highlight the override and remove it (hit the delete key).
If you’re making the adjustments to your Xcode project directly, be sure to select the “Project” instead of the “Target” on the left pane.

iPhone 5 Performance - Fastest Smartphone

According to Anandtech.com, iPhone processor, A6 is an ARM-compatible system-on-a-chip, ARMv7 based, designed entirely by Apple. PCMag.com tested Web browser performance of A6 and its system performance, Then they were compared with previous iPhone and current Android phone, for the browser benchmarks: Sunspider, Browsermark, and Guimark 3 Bitmap. Sunspider is about JavaScript, Guimark is about interactive HTML5, and Browsermark is an overall browser benchmark. 

Phones
Processors
iPhone  5
A6
iPhone  4S
Dual-core 1G (underclocked to 800MHz) A5
iPhone  4
1GHz (underclocked to 800MHz) A4
iPhone 3GS
833 MHz (underclocked to 600 Mh) z ARM Cortex-A8
iPhone / iPhone 3G
620 MHz (underclocked to 412 MHz) ARM 1176JZ(F)-S
Samsung Galaxy S III
1.5GHz Qualcomm Snapdragon S4 (Dual – core Krait)
Motorola Droid RAZR M
1.5GHz Qualcomm Snapdragon S4 (Dual – core Krait)

The followings are PCMag.com's results:

chart

iPhone 5 vs Android Chart

Sunday, September 16, 2012

Sony’s HMZ-T2 Head-mounted 3D Display

Sony revealed an incredible new head-mounted 3D display visor, the HMZ Immersive Pleasure Device, the HMZ-T2, Sony is going to be showing the visor at this year’s Tokyo Game Show, with public trials held on September 22nd and 23rd.

The original model, the HMZ T1, released last year, was a surprise hit for Sony, which has been criticized for a lack of innovative products as it goes through a prolonged restructuring.

The HMZ T2, the second edition of the device, is over 20 percent lighter than its predecessor and provides 5.1-channel virtual sound. The headphones can also be swapped out, allowing users to listen via their own equipment.

It features two OLED panels running at a resolution of 1280×720, with 3D immersion. The headset has two inner-ear speakers that provide virtual surround sound, and the unit only weighs 330 grams (11.6 ounces), but that’s before you count the 600 gram (21 gram) power supply unit. Sony has only announced the HMZ-T2 for launch in Japan at this point, hitting shelves on October 13th with a wallet-busting price of 70,000 yen (approx. $900).


Sunday, September 9, 2012

DS28CN01 1Kbit I²C/SMBus EEPROM with SHA-1 Engine

The DS28CN01 combines 1024 bits of EEPROM with challenge-and-response authentication security implemented with the Federal Information Publications (FIPS) 180-1/180-2 and ISO/IEC 10118-3 Secure Hash Algorithm (SHA-1). The memory is organized as four pages of 32 bytes each. Data copy-protection and EPROM emulation features are supported for each memory page. Each DS28CN01 has a guaranteed unique factory-programmed 64-bit registration number. Communication with the DS28CN01 is accomplished through an industry standard I²C- and SMBus™-compatible interface. The SMBus timeout feature resets the device's interface if a bus-timeout fault condition is detected. It can be used to protect your board design against piracy.


The working Mechanism


With DS28CN01, the MAC computation incorporates a random challenge chosen by the MAC recipient. Figure 1 illustrates the general concept. The longer the challenge, the more difficult it is to record all possible responses for a potential replay.

Figure 1. The challenge-and-response authentication process proves the authenticity of a MAC originator.
Figure 1. The challenge-and-response authentication process proves the authenticity of a MAC originator.

To prove the authenticity of the MAC originator, the MAC recipient generates a random number and sends it as a challenge to the originator. The MAC originator must then compute a new MAC based on the secret key, the message, and the recipient's challenge. The originator then sends the computed result back to the recipient. If the originator proves capable of generating a valid MAC for any challenge, it is very certain that it knows the secret and, therefore, can be considered authentic. The technical term for this process is challenge-and-response authentication. 




 


Diagram

DS28CN01: Typical Operating Circuit
Typical Operating Circuit

Technical Documents


Evaluation Kits

DS28CN01EVKITEvaluation Board/Evaluation System for the DS28CN01
DSAUTHSKEvaluation Kit for the DS28E01-100, DS28CN01, and DS2460

Blackmagic Announces 2nd Model of Cinema Camera with Passive Micro Four Thirds

timeless.jpg

Blackmagic Design announced a second model of the Blackmagic Cinema Camera that features passive Micro Four Thirds lens (MFT) mount which supports any Micro Four Thirds with manual iris and focus, and is also easily adapted to other lens mounts such as PL via third party adapters. The ability to use third party adapters to allow other types of lens mount is due to the Micro Four Thirds lens mount being much closer to the image sensor and allowing space for adapters to other lens mounts.
blackmagic cinema camera mft
It features
  • Blackmagic Cinema Camera MFT is identical to the Blackmagic Cinema Camera EF model.
  • High resolution 2.5K sensor allows improved anti aliasing and reframing shots.
  • Super wide 13 stops of dynamic range allows capture of increased details for feature film look.
  • Built in SSD allows high bandwidth recording of RAW video and long duration compressed video.
  • Open file formats compatible with popular NLE software such as CinemaDNG 12 bit RAW, Apple ProRes and Avid DNxHD. No custom file formats.
  • Includes no custom connections. Standard jack mic/line audio in, BNC 3 Gb/s SDI out, headphone, high speed Thunderbolt I/O technology, LANC remote control and standard DC 12-30V power connection.
  • Capacitive touch screen LCD for camera settings and “slate” metadata entry.
  • Compatible with extremely high quality Micro Four Thirds lenses and other mounts via adapters.
  • Supports 2.5K and 1080HD resolution capture in 23.98, 24, 25, 29.97 and 30 fps.
  • Thunderbolt connection allows direct camera capture via included Media Express software and supports live waveform monitoring via the included Blackmagic UltraScope software.
  • Includes a full copy of DaVinci Resolve 9.0 color grading software.
Blackmagic Cinema Camera MFT will be available December 2012 for US$2,995 from Blackmagic Design resellers worldwide.
metadata.jpg

FBI launches $1 billion face recognition project

Recently-released documents show that the FBI has been working since late 2011 with four states—Michigan, Hawaii, Maryland, and possibly Oregon—to ramp up the Next Generation Identification (NGI) Facial Recognition Program. When the program is fully deployed in 2014, the FBI expects its facial recognition database will contain at least 12 million “searchable frontal photos. According to New Scientist, this program will cost $1 billion.
biometrics1.jpg
Tests in 2010 showed that the best algorithms can pick someone out in a pool of 1.6 million mugshots 92 per cent of the time. It's possible to match a mugshot to a photo of a person who isn't looking at the camera too. Algorithms such as one developed by Marios Savvides's lab at Carnegie Mellon can analyse features of a front and side view set of mugshots, create a 3D model of the face, rotate it as much as 70 degrees to match the angle of the face in the photo, and then match the new 2D image with a fairly high degree of accuracy. The most difficult faces to match are those in low light. Merging photos from visible and infrared spectra can sharpen these images, but infrared cameras are still very expensive.

facial2.jpg
According to JENNIFER LYNCH,  "FBI Ramps Up Next Generation ID Roll-Out—Will You End Up in the Database",  one of the FBI’s goals for NGI is to be able to track people as they move from one location to another. Recent advancements in camera and surveillance technology over the last few years will support this goal. For example, in a National Institute of Justice presentation (pdf, p.17) at the same 2010 biometrics conference, the agency discussed a new 3D binocular and camera that allows realtime facial acquisition and recognition at 1000 meters. The tool wirelessly transmits images to a server, which searches them against a photo database and identifies the photo's subject. As of 2010, these binoculars were already in field-testing with the Los Angeles Sheriff’s Department. Presumably, the backend technology for these binoculars could be incorporated into other tools like body-mounted video cameras or the MORIS (Mobile Offender Recognition and Information System) iPhone add-on that some police officers are already using.
NIJ Facial Recognition Binoculars
Private security cameras and the cameras already in use by police departments have also advanced. They are more capable of capturing the details and facial features necessary to support facial recognition-based searches, and the software supporting them allows photo manipulation that can improve the chances of matching a photo to a person already in the database. For example, Gigapixel technology, which creates a panorama photo of lots of megapixel images stitched together (like those taken by security cameras), allows anyone viewing the photo to drill down to see and tag faces from even the largest crowd photos. And image enhancement software, already in use by some local law enforcement, can adjust photos"taken in the wild" (pdf, p.10) so they work better with facial recognition searches. 
Cameras are also being incorporated into more and more devices that are capable of tracking Americans and can provide that data to law enforcement. For example, one of the largest manufacturers of highway toll collection systems recently filed a patent application to incorporate cameras into the transponder that sits on the dashboard in your car. This manufacturer's transponders are already in 22 million cars, and law enforcement already uses this data to track subjects. While a patent application does not mean the company is currently manufacturing or trying to sell the devices, it certainly shows they're interested.

Followers

Blog Archive

About Me

My photo
HD Multimedia Technology player