| Proj No. | A2011-251 |
| Title | Silicon FET Reservoir for Dynamic Edge Vision-3 |
| Summary | The main aim of edge AI is to offer real time inference of data and has important applications in autonomous navigation. However, because of its limited computational resources, strategies are required for managing the collection and preprocessing of data to prevent the voluminous data from overwhelming the machine learning model. In the realm of computer vision, significant progress has been made in terms of a silicon retina sensor or event camera to reduce the visual data sampled. However, this device outputs a time-series data which require complex models such as a recurrent or spiking neural network for analysis. Therefore, to fully leverage the advantage of the event camera, innovative solutions are needed to preprocess the time-based event stream so that conventional well-established algorithm already formulated for static images can be deployed. In this regard, our team has developed a FET reservoir that can compress an event stream into a static frame without loss of temporal information. Using our developed FET reservoir simulation model, the FYP student will investigate its performance using different datasets such as DVS128, DVSLip, SL-Animals-DVS, etc. We seek passionate students who have background in Python programming. |
| Supervisor | A/P Ang Diing Shenp (Loc:S2 > S2 B2C > S2 B2C 95, Ext: +65 67906023) |
| Co-Supervisor | - |
| RI Co-Supervisor | - |
| Lab | Semiconductor Characterizatio (Loc: S1-B3C-27A) |
| Single/Group: | Single |
| Area: | Smart Electronics and IC design |
| ISP/RI/SMP/SCP?: |