M.Sc Thesis

M.Sc StudentCorem Gila
SubjectSupport for Novice Users in Computer-Aided Design Using
Natural User Interface
DepartmentDepartment of Architecture and Town Planning
Supervisors ASSOCIATE PROF. Noemi Bitterman
PROF. Gershon Elber
Full Thesis textFull thesis text - English Version


This work describes a design process of a Natural User Interface (NUI) system for a 3D modeling application. The theoretical model supporting NUI and state-of-the-art

technologies and projects - commercial and academic are described at the beginning of the work.

A CAD tools survey was held and revealed the main differences between designs aimed for novice and expert users. Conclusions from this part are used in the systems design that follow.

A survey of gestures that studies users’ mental models regarding gestures that describe functions found in any typical Computer-Aided Design (CAD) application. Results show that the mental model for manipulations such as Move, Rotate and Select are similar amongst most users as opposed to functions such as Undo / redo or Display Side / Top View of a model that yield a large number of options, suggesting a lack of a concrete common mental  model.

Following the survey of gestures, the system was designed and implemented via several  sub-sections: (a) design goals and scenarios, (b) Evaluation of gestures models using the House of Quality method, (c) System design and (d) Evaluation of the system via user observations.

Finally, a revised model was suggested. The new model is based on the results from the  system design process and study of gestures with the goal of minimizing functions and simplifying interaction. The model is represented via a low fidelity method - short  animations that show hand gestures triggering CAD functions as well as relevant interaction.

Results from a survey of these functions, show that the revised model is highly remembered  and recalled. Results also validate findings from the survey of gestures, i.e., manipulations,  being direct and deeply rooted in our cognitive system, are easier to remember than  gestures that are indirect and symbolic actions (see section 1.5 for a discussion about the  differences between gestures and manipulations).

As a conclusion we suggest a wide use of manipulations when a NUI system is designed, while gestures should be used limitedly and only after proper user testing to confirm these gestures are part of the users’ cognitive system.