基于全光电忆阻器的储备池系统的动作分类方法
By implementing multifunctional and multimodal optoelectronic devices in the reservoir system using all-photon memristors, the problem of complex hardware configuration in reservoir computing is solved, and high-precision classification of video motion and improved device integration density are achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- TSINGHUA UNIVERSITY
- Filing Date
- 2023-05-25
- Publication Date
- 2026-07-17
AI Technical Summary
The existing reservoir computing requires the configuration of two different hardware components, resulting in high architectural and integration complexity. Furthermore, the separation of sensors, processors, and storage in traditional artificial vision systems leads to redundant information exchange and increased energy consumption, which limits device integration density and processing time.
Design a storage pool system based on all-opto-memristors. By realizing a multifunctional multimodal opto-device that integrates sensing, storage and computing on the same hardware, and utilizing the dynamic opto-mode and non-volatile opto-mode of all-opto-memristors, achieve high-precision classification of video motion.
The system integrates sensing, storage, and computing on the same hardware, improving device integration density and processing time, and enabling high-precision classification of video actions.
Smart Images
Figure CN116883713B_ABST