Images generated by projecting shape information in three-dimensional space onto a plane, and/or a series of these technologies.
Software and system that enable computers to make human-like perceptions and judgments such as computer programs that understand and judge sentences, images, conversations, and sounds. AI stands for Artificial Intelligence.
Formulated procedure for solving a problem in mathematics, computing, linguistics, or related fields.
A type of machine learning method that realizes artificial intelligence by utilizing human brain imitated neural network mechanism, which is being commercialized in the field of image recognition.
A computer system that is built into such as home appliances and machines to achieve specific functions.
A company that does not own a fab (factory) and manufactures, or such a business model. DMP is also a research and development fabless semiconductor vendor that does not have its own factory.
A method to increase the speed in parallel computing part of the application by utilizing GPU to calculate that was conventionally processed by CPU alone. Depending on the target application, for example, in the case of video transcoding, the speed can be several times to 100 times faster than when calculating by CPU alone. GPGPU stands for General-Purpose computing on Graphics Processing Unit.
Arithmetic unit or processor specialized in real-time image processing represented by computer games. GPU stands for Graphics Processing Unit.
A mechanism in which various “things” in society not limited to information communication devices such as personal computers, smartphones / tablets, and game consoles are connected to the Internet and control each other by exchanging information, realizing automatic recognition, automated control, and telemetry.
IoT stands for Internet of Things.
Partial circuit modules within an LSI, designed for a specific function (e.g. graphics IP core). IP stands for Intellectual Property.
The most popular international standard for quality management systems in the world, used by over 1 million organizations in more than 170 countries. DMP acquired ISO 9001 certification in May 2019.
Large-scale integrated circuits composed of silicon wafers (materials with properties intermediate between conductors and insulators used in the manufacture of semiconductor products). LSI stands for Large Scale Integration and is also called “semiconductor”.
DMP’s proprietary patented technology that realizes high-performance graphics processing and low power consumption by converting the standard processing system used in the programmable shader architecture into hardware. It is used in Nintendo’s portable game consoles and so on.
A trademark for a graphics IP core that is compatible with international standards and can be equipped with DMP’s proprietary extended function “MAESTRO”.
DMP’s graphics processor for next-generation entertainment jointly developed with BANDAI NAMCO Entertainment Inc. It integrates 2D video and 3D graphics functions at a high level and has specifications optimized for amusement machines.
Integrated circuit (design method) that integrates a series of functions (systems) required on one semiconductor chip. SoC stands for System on a Chip.
An AI FPGA module that provides high-performance AI inference processing and power saving and fan-lessness provided by DMP (C2 for Intel® version, C3 for Xilinx version). DMP also provides “ZIA C2/C3 KIT” as a development kit for system studies.
An image recognition and image classification engine that uses AI (artificial intelligence), developed independently based on the image processing technology that DMP has cultivated so far.
An ultra-low power consumption processor IP provided by DMP that specializes in deep learning inference processing on the edge side and enables inference processing on data such as images, videos, and audio. DV720 is the successor to DV700, achieving 60% smaller size and 3.2 times higher performance (when mounted on Xilinx’s Zynq® UltraScale+ MPSoC).
DV500, like DV720, supports 16-bit floating point, and is optimized for the scene understanding and object recognition functions often used in industrial equipment and car driving support systems, enabling reduction in size while maintaining high processing performance.