VeriSilicon has launched a new AI image processing IP series to empower the era of high-definition imaging
In 2025On February 27, Shanghai – VeriSilicon (VeriSilicon, stock code: 688521. SH) today announced the launch of its latest generation of AI image processing IP series, including intelligent noise reduction IP (AINR1000 and AINR2000) and advanced super-resolution IP (AISR1000 and AISR2000). The IP provides efficient, flexible, and cost-effective solutions for automotive, surveillance, cloud gaming, and consumer electronics with superior performance, power, and area (PPA) characteristics.
VeriSilicon’s self-developed AI pixel processing algorithm empowers AINR and AISR series IPs to achieve excellent image processing capabilities. AINR1000 and AINR2000 effectively reduce image noise while precisely preserving detail and color fidelity, especially in low light or complex lighting conditions. AINR1000 supports 30 frames/second processing at 2560×1440 resolution, making it suitable for low-resolution applications such as entry-level webcams; AINR2000 supports 60 frames per second processing at 5600×4208 resolution to meet the needs of high-end security systems and action cameras.
AISR1000 and AISR2000 are equipped with advanced AI super-resolution technology, which can upscale SD, HD and FHD video sources to 4K resolution with high quality, and support multiple input sources such as real-time video, cloud gaming and cloud desktop. Significantly enhance image details and eliminate aliasing through innovative operator layer and layer fusion, resulting in clearer, more natural images. AISR1000 supports processing at up to 60 frames per second at up to 4K resolution, while the AISR2000 features a dual-engine architecture that achieves processing speeds of up to 144 frames per second at 4K resolution with optimized performance and power efficiency. The AI-SR series IPs support up to five times the flexible scaling scale, ensuring superior visual quality and scalability.
VeriSilicon’s FLEXA interface communication technology ensures that these IPs work with VeriSilicon’s ISPs, VPUs, and displaysprocessorSeamless IP integration allows for efficient and flexible embedding into existing imaging pipelines, improving overall system performance. At the same time, these IPs are also compatible with third-party pixel processing pipelines and are suitable for a variety of scenarios.
“Our AI-NR and AI-SR series IPs leverage cutting-edge AI technologies to significantly improve imaging quality while balancing low power consumption and high efficiency,” said Weijin Dai, Chief Strategy Officer, Executive Vice President and General Manager of VeriSilicon’s IP Division. These IPs are widely used in devices such as battery-powered cameras and mobile application processors, and the AI-SR series IP has been deployed in data centers to enhance video transcoding capabilities for AI-driven video applications. We are committed to providing comprehensive image enhancement functions to help customers develop high-performance products, and continue to expand the layout of AI-based image preprocessing and post-processing technologies to continuously improve image quality and efficiency. ”
VeriSilicon’s AI-NR and AI-SR series IPs are now available for licensing and integration. For more information and to schedule a demo, please visit the relevant link.
The above is the detailed content of VeriSilicon’s launch of low-power AI noise reduction and AI super-resolution series IP, for more information, please pay attention to other related articles on php Chinese website!
