Description
Open-Source Silicon at Stanford (OS3) has partnered with a lab from Stanford Medicine to improve the way surgeons analyze and assess surgery videos in real time. Currently, the video data is processed on a remote server using GPUs, which introduces a delay that can affect surgical outcomes. OS3 plans to develop technology that processes raw video data locally, significantly reducing this delay. This improvement is vital in medical projects where even a fraction of a second can make a difference.
The Future: Real-Time Micro-Inferencer Unit
The project focuses on developing the micro-inferencer unit, which uses application-specific integrated circuits (ASIC) and field-programmable gate arrays (FPGA) to handle various data types like video and sensor data. The μInferencer processes this data to produce outputs such as video encoding, segmentation, and pulsality, aiming to minimize delay. This advance in real-time data processing has applications in medical diagnostics and autonomous vehicles.
Timeline
- Project Scope Defined: June 25, 2024
- Design Phase Begins: July 1, 2024
- Tapeout Deadline for SKY130: November 11, 2024
- SKY130 Tapeout Back from Fab: April 11, 2025
- TSMC 65 Shuttle: Spring 2025