| Proj No. | A2087-251 |
| Title | Synthesis PCB Image Analysis and Quality Assessment |
| Summary | Recent generative techniques can create realistic PCB images, but determining how closely they match actual PCBs remains challenging. The project involves using detection and classification models to compare key features—such as component arrangement, trace routing, and design details—between synthetic images and ground truth data. By analyzing these differences, the project will develop quantitative metrics to measure the realism and accuracy of the synthetic images. This framework not only improves the evaluation of generated PCB images but also supports more reliable automated inspection and design verification in PCB manufacturing. Moreover, it will leverage advanced machine learning methods to identify subtle discrepancies, thereby enabling ongoing refinement of generative models and ensuring that synthetic datasets accurately reflect real-world PCB designs. |
| Supervisor | A/P Gwee Bah Hwee (Loc:S1 > S1 B1B > S1 B1B 42, Ext: +65 67906861) |
| Co-Supervisor | - |
| RI Co-Supervisor | - |
| Lab | IC DESIGN II (Loc: S1-B2B-10) |
| Single/Group: | Single |
| Area: | Intelligent Systems and Control Engineering |
| ISP/RI/SMP/SCP?: |