Artificial Intelligence-Enabled Marking for Computer-Aided Design Drawings

Artificial Intelligence-Enabled Marking for Computer-Aided Drawings (AMCAD) is the first and only automated system on the market capable of grading CAD exercises accurately. It combines two methods in computer vision technology to view and detect differences between students’ drawings and the answer key, independently of their position and size. Existing CAD auto-grading systems either assess unreliable attributes, superimpose drawings indiscriminately, or generate distortions. This web application circumvents manual marking of CAD drawings, which is extremely time-consuming and prone to human errors. The application, in turn, allows students to attempt more exercises, as lecturers can provide prompt feedback to them. Not all CAD automated marking systems provide post-comparison scores. Those that do are rubrics-based and not helpful in highlighting to students the errors in their drawings. AMCAD assigns a score for best-matched views against best drawing alignment and highlights drawing errors using superimposition or side-by-side comparisons.

Innovators

• Lim See Yew, Senior Director, Engineering
• Lin Kok On, Course Manager, Marine Engineering
• Tan Chee Keong, Senior Lecturer, Precision Engineering-Technical Engineer Diploma
• Agnes Tan Hwee Juan, Lecturer, Mechanical Engineering
• Ricky Tan Yuan Sheng, Lecturer, Precision Engineering-Technical Engineer Diploma
• Lisa Wang Jianwu, Senior Lecturer, Built Environment