VeriSilicon has unveiled AcuityPercept, an innovative AI-powered automatic Image Signal Processing (ISP) tuning system aimed at optimizing image processing parameters for enhanced object recognition.

By continuously refining ISP settings, AcuityPercept significantly boosts object detection accuracy and delivers high-quality images suitable for neural network processing.

This dynamic optimization of ISP settings enhances the accuracy and efficiency of AI perception systems across various sectors, including autonomous driving and robotics.

VeriSilicon is dedicated to offering comprehensive custom silicon services and semiconductor IP licensing, leveraging its proprietary semiconductor technology.

Weijin Dai, Chief Strategy Officer at VeriSilicon, underscored the critical role of AcuityPercept in integrating ISP with AI perception engines, particularly for improving object recognition in vital applications such as smart surveillance and autonomous vehicles.

AcuityPercept has already demonstrated promising results with data from automotive ISP customers, showcasing its ability to reliably identify objects in diverse environments.

The system employs a sophisticated global directive and local refinement algorithm, which processes metadata and feedback from AI task models to achieve optimal ISP tuning.