videantis GmbH is a semiconductor intellectual property (IP) and software design company headquartered in Hannover, Germany, and operating as a subsidiary of Mythic. Founded in 2004 by Andreas Dehnhardt, the company specializes in developing energy-efficient, unified processor architectures and software solutions tailored for deep learning, computer vision, multi-modal sensor fusion, image processing, and multi-format video coding.
Videantis silicon IP has been deployed extensively in tier-1 automotive manufacturing, smart consumer electronics, robotics, edge computing, and industrial surveillance systems, with deployments spanning tens of millions of commercial vehicles.
Core Technology and Architecture
The cornerstone of videantis's offerings is its unified multi-core processor architecture. Rather than relying on separate, heterogeneous hardware accelerators for different tasks—such as dedicated video codecs, separate DSPs for computer vision, and distinct neural processing units (NPUs)—videantis utilizes a programmable architecture capable of executing diverse algorithms on a single unified platform.
Key architectural characteristics include:
Unified Compute Engines: Enables simultaneous execution of convolutional neural networks (CNNs), classical computer vision primitives, digital signal processing (DSP), and standard video compression/decompression on identical compute units.
Scalability and Modularity: Core configurations scale from low-power single-core sub-systems for edge sensors to high-throughput multi-core clusters designed for central compute and autonomous vehicle domain controllers.
Energy and Area Efficiency: Designed for aggressive silicon area reduction and low thermal footprints, facilitating integration into battery-powered devices and thermally constrained automotive electronic control units (ECUs).
Functional Safety and ASIL Compliance: Supports fail-operational processing architectures, hardware redundancy, and diagnostics up to ISO 26262 ASIL D standards for automotive safety-critical applications.
Products and Intellectual Property
1. Processor IP Cores
v-MP6000UDX Processing Platform: The company's flagship unified processor IP core family, optimized for heavy-load AI inference, advanced computer vision, and multi-stream video encode/decode. It handles full-resolution sensor preprocessing, multi-camera streaming, and deep neural network layers natively.
Scalable Multi-Core Subsystems: Configurable hardware designs allowing semiconductor manufacturers to integrate clusters of videantis processing cores connected via coherent on-chip interconnects.
2. Video Compression and Coding IP
Multi-Format Codec Support: Hardware/software implementations supporting leading video codecs, including H.266/VVC, H.265/HEVC, H.264/AVC, VP9, AV1, and legacy formats.
Low-Latency Streaming: Specialized pipelines designed for real-time video transmission in automotive surround-view systems, electronic mirrors (CMS), and remote drone control.
3. Software Tools and Development Environments
VIVID (videantis Vision Integrated Development): An integrated software suite and toolflow designed to streamline application porting and optimization.
Neural Network Toolflow: Automates the quantization, pruning, compilation, and mapping of industry-standard neural network frameworks (such as PyTorch, ONNX, and TensorFlow) directly to the unified processor cores.
Computer Vision and DSP Libraries: Optimized software routines adhering to APIs such as OpenVX and OpenCV for image enhancement, feature extraction, filtering, optical flow, and depth mapping.
Automotive-Grade Software Stacks: Automotive SPICE (ASPICE) Level 2 certified software stacks for integration with AUTOSAR environments and functional safety requirements.
Services Offered
Semiconductor IP Licensing: Licensing of synthesizable RTL (Register-Transfer Level) processor designs to fabless chip companies, integrated device manufacturers (IDMs), and Tier-1 automotive suppliers.
Custom Architecture Tailoring: Tailoring core counts, cache hierarchies, and bus interfaces to match specific client SoC constraints and power budgets.
Algorithm Porting and Optimization: Custom software porting of proprietary customer vision pipelines, custom neural network layers, or proprietary codecs to the videantis hardware platform.
System-on-Chip (SoC) Integration Support: Technical consultation covering synthesis, place-and-route verification, silicon bring-up, and production test generation.
Target Markets and Applications
Automotive and Autonomous Driving (ADAS/AD): Front cameras, driver monitoring systems (DMS), occupant monitoring systems (OMS), surround-view monitors, radar/LiDAR sensor fusion, and automated parking systems.
Smart Surveillance and Smart Cities: Real-time facial recognition, license plate reading, edge-based object classification, and privacy redaction in network video recorders (NVRs) and IP cameras.
Consumer Electronics and XR: Augmented reality (AR) and virtual reality (VR) tracking, gesture interfaces, action cameras, and smart home appliances.
Robotics and Industrial Automation: Vision-guided robotic navigation (SLAM), automated visual inspection, and aerial drone situational awareness.
Please note: This information is generated by AI. For legal use, verify with company sources.