No abstract

Understanding and tailoring the visual elements of a developing product to evoke desired perceived qualities and a positive response from the consumer is a key challenge in industrial design. To date, computational approaches to assist this process have either relied on stiff geometric representations, or focused on superficial features that exclude often elusive shape characteristics. In this work, we aim to study the relationship between product geometry and consumers' qualitative judgments through a visual decomposition and abstraction of existing products. At the heart of our investigation is a shape analysis method that produces a spectrum of abstractions for a given three-dimensional (3D) computer model. Our approach produces a hierarchical simplification of an end product, whereby consumer response to geometric elements can be statistically studied across different products, as well as across the different abstractions of one particular product. The results of our case study show that consumer judgments formed by coarse product “impressions” are strongly correlated with those evoked by the final production models. This outcome highlights the importance of early geometric explorations and assessments before committing to detailed design efforts.

No abstract

Problems faced by engineering students involve multiple pathways to solution. Students rarely receive effective formative feedback on handwritten homework. This paper examines the potential for computer-based formative assessment of student solutions to multipath engineering problems. In particular, an intelligent tutor approach is adopted and tested out on problems of truss analysis, studied in engineering statics. With a cognitive model for solving the class of problems, the tutor allows the student wide latitude in solution steps, while maintaining sufficient constraints for judging the solution and offering feedback. Proper selection of judging points prevents interference with productive student work, while avoiding accumulated errors. To monitor student learning, efforts to apply distinct skills were extracted on the fly from student work. Using statistical methods developed for intelligent tutoring systems, metrics of the effectiveness of the feedback and areas for further improvements were gleaned from error rates in successive opportunities to apply distinct skills.

Hand-drawn sketches are powerful cognitive devices for the efficient exploration, visualization and communication of emerging ideas in engineering design. It is desirable that CAD/CAE tools be able to recognize the back-of-the-envelope sketches and extract the intended engineering models. Yet this is a nontrivial task for freehand sketches. Here we present a novel, neural network-based approach designed for the recognition of network-like sketches. Our approach leverages a trainable, detector/recognizer and an autonomous procedure for the generation of training samples. Prior to deployment, a Convolutional Neural Network is trained on a few labeled prototypical sketches and learns the definitions of the visual objects. When deployed, the trained network scans the input sketch at different resolutions with a fixed-size sliding window, detects instances of defined symbols and outputs an engineering model. We demonstrate the effectiveness of the proposed approach in different engineering domains with different types of sketching inputs.