Toshiba Launches New Line-up of Image Recognition Processors for Monocular Cameras

- Enables Night-Time Pedestrian Detection in ADAS Applications -

TOKYO — (BUSINESS WIRE) — February 23, 2016Toshiba Corporation (TOKYO:6502) today announced that it will expand its TMPV760 series line-up of image recognition processors for automotive applications with the launch of “TMPV7602XBG”, which is suited to monocular cameras in the Advanced Driver Assistance Systems (ADAS) applications. Sample shipments of TMPV7602XBG will start in March 2016, with mass production scheduled for January 2018.

This Smart News Release features multimedia. View the full release here: http://www.businesswire.com/news/home/20160223005850/en/

Toshiba: "TMPV7602XBG", a new TMPV760 series line-up of image recognition processors suited to monoc ...

Toshiba: "TMPV7602XBG", a new TMPV760 series line-up of image recognition processors suited to monocular cameras in ADAS applications. (Photo: Business Wire)

The TMPV7602XBG provides high-speed, low-power image recognition of images captured by an on-vehicle monocular camera. It recognizes information necessary for safe driving, such as nearby pedestrians, vehicles, traffic lanes, traffic signs, road markings, and traffic lights.

The European New Car Assessment Programme (Euro NCAP) has released an updated roadmap for safety ratings, and the section on test items requires adoption of the following items by 2018: “pedestrian detection and collision avoidance systems in daylight and obscure lighting conditions” and “pedal cyclist detection and collision avoidance systems”. Toshiba’s new image recognition processor targets ADAS applications that need to cope with these new requirements.

While inheriting the high night-time object recognition performance of the TMPV760 series, TMPV7602XBG achieves low-power consumption and package downsizing (17.0mm×17.0mm, 521 balls, and BGA0.65mm pitch) by optimizing the circuits providing a single video input channel for monocular cameras. The TMPV7602XBG helps to inhibit heat generation in the on-vehicle camera and also contributes to a reduced package size.

ADAS, which includes autonomous braking and lane-keeping assistance, is now widely implemented in vehicles, from compact to luxury cars. Toshiba will continue to develop image recognition processors that deal with increasing deployment of systems for safe driving support.

Key Features of the New Product

1. Higher Recognition Accuracy, Especially at Night
The TMPV7602XBG inherits the high night-time object recognition performance of the TMPV760 series. Toshiba’s color-based feature descriptor, the Enhanced CoHOG[1] accelerator, is integrated into the TMPV7602XBG and supports four sets of feature descriptors[2], such as the image feature descriptor CoHOG and color-based image feature descriptor Color-CoHOG. This provides higher recognition accuracy [3], especially at night and at times of low luminance difference between objects and the background.

2. Downsizing the Camera Module by Miniaturizing a Package and Reducing Power Consumption
Package size of the TMPV7602XBG has been reduced (17.0mm×17.0mm, 521 balls, and BGA0.65mm pitch) by optimizing the circuits to downsize the ADAS monocular camera module. Its power consumption is expected to be 50% [4] lower than that of the existing TMPV760 series product, TMPV7608XBG. In addition to this, the newly-incorporated power management unit allows efficient power management by controlling each functional block when it switches to or from sleep mode.

[1] CoHOG is Toshiba’s original Co-occurrence Histograms of Oriented Gradients technology.
[2] Color-CoHOG, Co-occurrence Histograms of pairs of Edge orientations and color Differences (CoHED), Co-occurrence Histograms of color Difference (CoHD), and Color Histogram.
[3] It is equivalent to the day-time recognition rate of the TMPV75 Series.
[4] Toshiba’s data of the image processing are being collected for product development.

Outline of New Product

Part number   Sample price

(tax included)

  Sample shipment   Mass production
TMPV7602XBG   7,000 yen   March 2016   January 2018
 

Main Specifications

Part number   TMPV7602XBG
CPU Toshiba original 32-bit RISC CPU

Media embedded Processor(MeP) × 2

Image recognition engine Toshiba original 32-bit RISC

Media Processing Engine(MPE) × 4

Image
processing
accelerators

  Affine transform 2 ch
Filter 2 ch
Histogram 2 ch

Histogram of Oriented
Gradients

2 ch
(CoHOG/HOG)

Enhanced CoHOG 1 ch
Matching 2 ch
Pyramid 2 ch

On-chip
peripheral
functions

Video input interface 1 ch
Video output interface   1 ch

Power supply
voltage

Core: 1.1V, I/O: 3.3V and 1.2V (connect LPDDR2), 1.2V (MIPI CSI2-I/F)

Operating
frequency

Max. 266.7MHz
Package   PLFBGA 521 balls, 17mm × 17mm, 0.65mm ball pitch

1 | 2  Next Page »
Featured Video
Editorial
Jobs
Manufacturing Test Engineer for Google at Prague, Czechia, Czech Republic
Mechanical Engineer 2 for Lam Research at Fremont, California
Senior Principal Mechanical Engineer for General Dynamics Mission Systems at Canonsburg, Pennsylvania
Equipment Engineer, Raxium for Google at Fremont, California
Mechanical Test Engineer, Platforms Infrastructure for Google at Mountain View, California
Mechanical Manufacturing Engineering Manager for Google at Sunnyvale, California
Upcoming Events
Celebrate Manufacturing Excellence at Anaheim Convention Center Anaheim CA - Feb 4 - 6, 2025
3DEXPERIENCE World 2025 at George R. Brown Convention Center Houston TX - Feb 23 - 26, 2025
TIMTOS 2025 at Nangang Exhibition Center Hall 1 & 2 (TaiNEX 1 & 2) TWTC Hall Taipei Taiwan - Mar 3 - 8, 2025
Additive Manufacturing Forum 2025 at Estrel Convention Cente Berlin Germany - Mar 17 - 18, 2025



© 2024 Internet Business Systems, Inc.
670 Aberdeen Way, Milpitas, CA 95035
+1 (408) 882-6554 — Contact Us, or visit our other sites:
AECCafe - Architectural Design and Engineering EDACafe - Electronic Design Automation GISCafe - Geographical Information Services TechJobsCafe - Technical Jobs and Resumes ShareCG - Share Computer Graphic (CG) Animation, 3D Art and 3D Models
  Privacy PolicyAdvertise