Fig 1.
A simplified example of the divergence-convergence cycles of shape exploration for the conceptual design of a water jar.
The basic shape can be progressively refined with rough sketches before developing the concept further. This work proposes an AI (Artificial Intelligence) tool, Shapi, for assisting shape exploration divergence as a process parallel to human sketching. Shapi generates variations around a given ‘Seed’, where Global variations can be used to catalyze insight or ideas about the overall shape or particular details anywhere in the contour, while Local variations can be used for more systematic inspections of the different shape regions.
Fig 2.
Drawing a blacktip shark, sketching as a reflective conversation.
The curves (top) are thicker and darker the more important they are for establishing and discovering the general character of the drawing.
Fig 3.
Suggested workflow: Using Shapi in a human-AI collaboration for early shape exploration.
Fig 4.
a) The general approach proposed to assist early shape divergence and exploration in a human-AI collaboration. Note that the human collaborator provides the Seed and the shape reinterpretation at the end of every cycle. b) An example shape starting with a given Seed and leading to suggested global and local variations with a rough sketch style.
Fig 5.
Examples of drawn boundaries and a symmetry axis to guide the exploration interactively.
Fig 6.
Explorations with different purposes.
Input Seeds (left) are explored with Shapi (middle), and depending on the interactive controls, this may inspire reinterpretations as product concepts (top) or entirely new entities with divergent emerging patterns (bottom).
Fig 7.
Some of the Global and Local shape variations generated with Shapi for the conceptual exploration of a hand-axe.
The Seed is based on an exploration by a product & communication design student [86].
Fig 8.
Some of the Global and Local shape variations generated with Shapi for the conceptual exploration of a coffee mug.
The Seed is based on the silhouette of a design by Dunoon [87].
Fig 9.
Some of the Global and Local shape variations generated with Shapi for the conceptual exploration of a soap dispenser.
The Seed is based on an exploration by product designer & illustrator Von der Heyde [88].
Fig 10.
Some of the Global and Local shape variations generated with Shapi for the conceptual exploration of a perfume bottle.
The Seed is based on an exploration by an industrial designer [89].
Fig 11.
Some of the Global and Local shape variations generated with Shapi for the conceptual exploration of a water bottle.
The Seed is based on an exploration by a product designer [90].
Fig 12.
Some of the Global and Local shape variations generated with Shapi for the conceptual exploration of a liquor bottle.
The Seed is based on an exploration by an industrial designer & entrepreneur [91].
Fig 13.
Some of the Global and Local shape variations generated with Shapi for the conceptual exploration of a lamp shade.
The Seed is based on an exploration by an industrial designer [92].
Fig 14.
Shapi variations of the silhouette of a car (center column) where some characteristics may be considered ‘emergent’ features.
On the right, some reinterpretations of how a designer might continue the ideation conversation. The Seed is based on an exploration by an interior photographer [93].
Fig 15.
Example of an output sheet for the Shapi Global Exploration of an abstract input Seed.
Fig 16.
Two abstract Seeds are explored with Shapi.
Several cycles of divergence-convergence result in the Shapi logo.
Fig 17.
Some implementations with artistic open-ended purposes.
The Seeds of three objects (hand-axe, lamp shade, and bottle) are used to reinterpret and discover new entities (human sketches on the right).
Fig 18.
Using an abstract Seed to generate new emerging patterns that inspire divergent human interpretations.
Fig 19.
Survey respondents by age and field of occupation.
Respondents may choose more than one occupation.
Fig 20.
Survey results of perceptions about Shapi’s capabilities and potential value with 0–5 scores.
Results are displayed by frequency, age group, occupation, and Postgraduate studies. Metrics include mean, median, Standard Deviation, Variation Ratio, and P-value.
Fig 21.
Survey results of perceptions about Shapi’s capabilities and potential value with 0–5 scores.
Results are displayed by frequency, age group, occupation, and Postgraduate studies.
Fig 22.
Word Clouds produced with the open feedback from the survey using all sentences (center), only the 20 most negative sentences (left), and the 20 most positive (right).
Fig 23.
Polarity of the open feedback from the survey.
It is displayed by frequency, age group, occupation, and Postgraduate studies.
Fig 24.
Correlations (r) and their P-values (P) comparing all distributions of perception scores and polarity.