In this thesis, we explored the use of generative artificial intelligence (AI) to visualize meat products, which are an essential part of a balanced daily diet. The research assessed the professional accuracy and visual appeal of AI-generated images compared to photographs taken with a camera, thereby evaluating the potential use of such visuals in the food industry.
In the first part of the experiments, images of ten chosen meat dishes and products were generated using an AI image generation tool, Adobe Firefly 3. To assess their professional adequacy, the generated images were presented to professionals in the field of food technology, who provided their opinions on the realism and accuracy of the depictions through a survey. After the first survey, photographs of meat products comparable to the generated images were taken. Finally, a survey was conducted among the general public, in which participants evaluated the appeal of the generated images and the photographs.
The results showed that generative AI has not yet achieved sufficient accuracy for the professional presentation of more complex meat products, as there are often inconsistencies in colour, texture, and structure. Nevertheless, the AI-generated images frequently appeared more aesthetically pleasing and were rated more attractive by the general public than the actual photographs.
In summary, generative AI already offers practical possibilities for food marketing, showing its potential, especially in promotional content, where professional precision is not critical. However, it also holds significant potential for development if generated images are to be used for professional purposes in the future. In such cases, we recommend using professionally accurate reference photographs.
|
