CAST Adds JPEG-LS Decoder to Lossless Image Compression IP Core Suite
CAST’s JPEG-LS Encoder and Decoder provide a scalable, zero-latency, ultra-low-power hardware solution for numerically- or visually-lossless image compression
WOODCLIFF LAKE, NJ USA -- June 12 2018 -- Semiconductor intellectual property provider CAST, Inc. has released a JPEG-LS Decoder IP core that—together with the JPEG-LS Encoder already available—provides an efficient hardware compression solution for the lossless or near-lossless transmission or storage of high quality images or video sequences.
The JPEG-LS cores provide:
- Superior compression ratios, yielding smaller files than PNG and smaller or similarly sized files than JPEG 2000 (see figure);
- Ultra-low power consumption, thanks to the low complexity of the JPEG-LS algorithm, which enables hardware implementations with an extremely small silicon footprint and minimal memory requirements;
- “Zero” (sub-millisecond) latency, because the line-based processing employed by JPEG-LS does not impose any buffering of frames or even image lines before the encoding or decoding process is initiated; and
- Scalable throughput, uniquely handling ultra-high resolutions and/or frame rates by internally combining and managing a configurable number of parallel processing engines.
Comparing the resulting file size and compression ratios of the PNG, JPEG 2000, and JPEG-LS lossless compression algorithms using typical images.
For example, consider the significant savings in hardware resources (16nm technology) possible for the lossless encoding and decoding of a 4K Ultra HD, 10-bit per color, RGB video stream at 30 frames per second:
JPEG 2000 | CAST JPEG-LS |
---|---|
several million gates several Mbits of memory more than 70mW tens of msec of latency | 350K gates 300 Kbits of SRAM less than 10mW less than 0.5 msec of latency |
Moreover, the CAST JPEG-LS encoder will yield a similar or even better compression ratio as the JPEG 2000 encoder, resulting in further resource savings system-wide.
Their competitive compression ratio with smaller hardware and lower latency make the JPEG-LS Encoder and Decoder cores an excellent choice for systems where artificial intelligence or video analytics process the video stream—such as for autonomous driving or augmented reality applications—because these systems typically do not accept image quality loss or any extra latency, while also requiring very low power consumption. Additional applications include those requiring high quality images with maximum file size savings, including space and medical imaging (JPEG-LS is part of the DICOM medical imaging standard).
About the JPEG-LS IP Cores
The JPEG-LS-E Encoderand JPEG-LS-D Decodereach fully support the JPEG-LS, ISO/IEC 14495-1 standard, which is based on the LOCO-I (LOw COmplexity LOssless COmpression for Images) algorithm developed by HP. The JPEG-LS-E has been used in products where decoding has been performed with standard software JPEG-LS decoders. The CAST Decoder appears to be the only such JPEG-LS decompression core commercially available; it can be used with the JPEG-LS-E or any other standard-conforming JPEG-LS encoder.
The JPEG-LS cores are easy to integrate and use, employing standard AXI-Stream and APB interfaces and operating standalone without processor intervention once programmed. Each is available in two versions, size-optimized, and scalable-throughput.
The single-pipeline, size-optimized versions process a single sample per clock cycle and require just one image line of buffering. The Encoder and Decoder can each handle several hundreds of Msamples per second when mapped to an ASIC technology; require as few as 40,000 ASIC gates; and run up to about 600 MHz and 350 MHz, respectively (on TSMC 16nm with SVT cells).
The multiple-pipe, throughput-optimized versions internally aggregate a user-specified number of the single-pipeline cores, handling multiple samples each cycle to efficiently process images or video with ultra-high resolutions and/or very fast frame rates. This scalable throughput capability appears to be unique to the CAST JPEG-LS cores.
The silicon-validated JPEG-LS Encoder and Decoder are available now, with royalty-free licensing for either ASICs (RTL) or FPGAs (netlists) from popular silicon providers.
About CAST
In addition to the JPEG-LS cores, CAST offers a range of compression solutions and image processing functions; 8051 microcontrollers and extreme-low-power 32-bit BA2X™ processors; industry-leading automotive interfaces; and a variety of peripherals, interfaces, security, and other IP cores. CAST IP features easy integration and reuse, royalty-free licensing, and availability for ASICs or FPGAs from all leading silicon providers. Learn more by visiting www.cast-inc.com, emailing info@cast-inc.com, or calling +1 201.391.8300.
|
CAST, Inc. Hot IP
Related News
- CAST Expands Image Compression Line with New JPEG-LS Encoder Core
- New PNG Encoder IP Core Expands CAST's Lossless Compression Suite
- CAST Adds Video and Image Processing Cores to Compression IP Product Line
- Alma Technologies Scales-up the Performance of the JPEG-LS Image Compression IP Core by Employing its Ultra High Throughput - UHT™ Architecture
- PNG Image Decoder IP Core Available from CAST and IObundle
Breaking News
- HPC customer engages Sondrel for high end chip design
- Ubitium Debuts First Universal RISC-V Processor to Enable AI at No Additional Cost, as It Raises $3.7M
- TSMC drives A16, 3D process technology
- Frontgrade Gaisler Unveils GR716B, a New Standard in Space-Grade Microcontrollers
- Blueshift Memory launches BlueFive processor, accelerating computation by up to 50 times and saving up to 65% energy
Most Popular
- Cadence Unveils Arm-Based System Chiplet
- Eliyan Ports Industry's Highest Performing PHY to Samsung Foundry SF4X Process Node, Achieving up to 40 Gbps Bandwidth at Unprecedented Power Levels with UCIe-Compliant Chiplet Interconnect Technology
- TSMC drives A16, 3D process technology
- CXL Fabless Startup Panmnesia Secures Over $60M in Series A Funding, Aiming to Lead the CXL Switch Silicon Chip and CXL IP
- Blueshift Memory launches BlueFive processor, accelerating computation by up to 50 times and saving up to 65% energy
E-mail This Article | Printer-Friendly Page |