Abstract: Computer science is experiencing rapid growth. Reconstructing cooking recipes from photos of food presents an intriguing challenge. The goal is to create complete recipes with ingredient lists, titles, and comprehensive instructions using convolutional layers in CNNs. About identifying complex patterns in food photos, this study clarifies the capabilities and limitations of CNNs by assessing critical performance metrics like recipe generation accuracy and efficiency. This work helps the culinary industry develop new technological solutions in response to the increasing need for a thorough understanding of meal preparation. Moreover, the implications of this research go beyond computer science, as it has the potential to drastically change how people interact with food. This research opens the door to a more diverse and interconnected culinary landscape by democratizing culinary knowledge and fostering a deeper understanding of global gastronomic traditions. In the end, this study's conclusions will guide future research into creating AI-powered culinary apps that are customized to each user's unique preferences and tastes, enhancing the appreciation of cooking around the world.

Keywords: Convolutional Neural Networks (CNNs), computer vision, culinary exploration, cooking instructions, deep learning architectures, Recipe1M dataset, image-to-recipe prediction, natural language processing, AI-driven culinary applications.

| DOI: 10.17148/IARJSET.2024.11589