Our Method: Distairation and Discussion
We have worked with Studio to cut the highest level of ross' personal sketches, use it to perform our Text-to-Image model, Image, Image. By training the model in the selected studio work, we were able to enter the main part of the Basic Language of Ross 'design, logical curves and organic patterns, allowing us to produce new ideas that focus on the unique ross' style.
From the beginning, we approached this work as an investigation of the people led by man. The study has decided to prioritize the language along the views of viewing data, working with and specify the Ross' Design Lexicon to direct the model effect. We focus on building a specific vocabulary that describes the studio work, knowing that the appropriate Refts are key to find reasonable results.
Throughout this process, Lovegrove Studio saw how the model responded to certain words and used these things to understand the targeted design results. This discussion between the Designer and AI was an important part of the process. Take care of how the model translates some words, using that diving response to promoting and directing from the studio vision. We challenge the model to produce a seat without using the word “seat,” using the factor to create a variety of results and the richest test of the form.
Result: Visible Protetype
We have developed many ideas with this special model and a team lovegrove Studio, then used Gemini to push the creative test to continue to explore the materials and visual view.
