Computational Technology Improves Safety Symbols
ISE Magazine – Volume 48: Number 02

Communicating hazard and precautionary information through warnings is an important part of industrial and product safety. researchers have found that many people poorly comprehend what safety warning pictorials mean. The stereotype production method commonly used to design symbols has potential users sketch pictorials and submit them to a professional designer. This designer then identifies similar themes or stereotypes among the sketches and produces final pictorial designs from those stereotypes. that poor comprehension of safety pictorials could result from an overreliance on a single individual, rather than the users themselves, when making final design decisions.

The sketch comparison and evaluation steps in the stereotype production method are particularly well-suited to mathematical clustering, while the synthesis of the final designs can be enhanced using interactive evolutionary computation. Two clustering algorithms were used to group these sketches into design families with similar attributes, resulting in five pictorial families and three primary graphical attributes: “pipe,” “lines of emission” and “flame.” The same primary attributes were found from both U.S. sketches and Indian sketches. design software application is under development that uses interactive evolutionary computation to propose and evaluate new pictorial designs based on these primary and secondary attributes.