Brother
Professional
- Messages
- 2,590
- Reaction score
- 539
- Points
- 113
Why pay designers if you have AI?
The success of a product in the market depends not only on its quality, but also on its compliance with the brand style. It is difficult for designers to create products that stand out from the competition, but at the same time maintain brand awareness.
Scientists from Carnegie Mellon University have presented a new technology that can make designers ' work easier. Experts have developed a fully automated neural network that can detect visual features associated with a brand. The model is called BIGNet (Brand Identification Graph Neural Network), and is trained on vector images of products.
BIGNet analyzes thousands of curves in the product image and highlights those that characterize the brand's style. Scientists note that before BIGNet, there was no method that would automatically extract style features using machine learning. Designers created rules in their heads, but they were difficult to formulate and difficult to communicate between products.
The team tested BIGNet on various products, such as popular Apple and Samsung smartphones. The model distinguished between the two brands with 100% accuracy, identifying specific features such as screen gaps and camera placement.
To demonstrate BIGNet's adaptability and generalizing ability across different products and design scales, the team evaluated BIGNet on 10 automotive brands. The model most accurately identified Audi, BMW and Mercedes-Benz, confirming that these manufacturers have a higher degree of style consistency than the budget car segment.
BIGNet features the design features of car brands
The technology can save designers significant time. Companies will no longer have to rely on people who have worked with them for more than 20 years to understand the brand's style.
Currently, BIGNet works with two-dimensional images taken directly. The researchers want to expand their tests on three-dimensional images and hope to develop a model to determine not only brand identity, but also other product characteristics. For example, what makes a car a muscle car or a sports car.
The development of BIGNet demonstrates how artificial intelligence can ensure accuracy and consistency in design, automatically identifying key elements that make a brand recognizable and attractive. The technology offers a significant improvement over traditional methods based on the subjective assessments and experience of designers, promising a new era in design where machines and human talent work together to create innovative and efficient products.
The success of a product in the market depends not only on its quality, but also on its compliance with the brand style. It is difficult for designers to create products that stand out from the competition, but at the same time maintain brand awareness.
Scientists from Carnegie Mellon University have presented a new technology that can make designers ' work easier. Experts have developed a fully automated neural network that can detect visual features associated with a brand. The model is called BIGNet (Brand Identification Graph Neural Network), and is trained on vector images of products.
BIGNet analyzes thousands of curves in the product image and highlights those that characterize the brand's style. Scientists note that before BIGNet, there was no method that would automatically extract style features using machine learning. Designers created rules in their heads, but they were difficult to formulate and difficult to communicate between products.
The team tested BIGNet on various products, such as popular Apple and Samsung smartphones. The model distinguished between the two brands with 100% accuracy, identifying specific features such as screen gaps and camera placement.
To demonstrate BIGNet's adaptability and generalizing ability across different products and design scales, the team evaluated BIGNet on 10 automotive brands. The model most accurately identified Audi, BMW and Mercedes-Benz, confirming that these manufacturers have a higher degree of style consistency than the budget car segment.
BIGNet features the design features of car brands
The technology can save designers significant time. Companies will no longer have to rely on people who have worked with them for more than 20 years to understand the brand's style.
Currently, BIGNet works with two-dimensional images taken directly. The researchers want to expand their tests on three-dimensional images and hope to develop a model to determine not only brand identity, but also other product characteristics. For example, what makes a car a muscle car or a sports car.
The development of BIGNet demonstrates how artificial intelligence can ensure accuracy and consistency in design, automatically identifying key elements that make a brand recognizable and attractive. The technology offers a significant improvement over traditional methods based on the subjective assessments and experience of designers, promising a new era in design where machines and human talent work together to create innovative and efficient products.
