Developing an efficient AI Vision involves various potential hardware and software blocks to detect, process, and instantly analyze thousands of visual objects, images, or actions in video images. Silicon and Sensor systems are evolved to achieve this efficient visioning using MIPI CSI2 sensor interfaces and AI/ML Accelerators in the form of a Neural Processing Unit (NPU). Let’s discuss how that implementation is handled in Linux Kernel by considering a real-time Object detection demo on Rockchip RK3588 NPU with Sony, IMX415 Camera sensor.
Here Amarula Solutions India Team @Jagan Teki @Suniel Mahesh is speaking about their implementation and work on this AI Vision Solutions.
