Ercan, G.B.Baran, M.Konca, E.Cetin, I.M.Korkmaz, I.2025-05-252025-05-25202597830318616111865-0929https://doi.org/10.1007/978-3-031-86162-8_12https://hdl.handle.net/20.500.14365/6207Netcetera and Ultra ComputingThis study aims to develop a robust system using image processing and machine learning to accurately differentiate poisonous and non-poisonous mushroom species, addressing the significant public health threat posed by poisonous mushroom consumption. Motivated by the urgent need for an efficient tool to aid mushroom enthusiasts, farmers, and healthcare professionals in real-time identification of harmful species, the research focuses on creating a mobile application capable of processing mushroom images, extracting pertinent features, and employing a well-trained machine learning model for precise toxic and non-toxic categorization. Through a diverse image dataset collection, preprocessing, feature extraction, and rigorous model evaluation, the study endeavors to enhance public safety and encourage the development of similar applications for species identification and environmental protection. Based on the experiments conducted, amongst many machine learning algorithms used to train a proper system to decide whether a mushroom is edible or poisonous, InceptionV3 deep learning model is chosen to be integrated into the mobile application implemented as the endpoint to the users. Additionally, a simple game is also embedded in the mobile app to make the users learn the poisonous mushrooms from their images. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.eninfo:eu-repo/semantics/closedAccessImage ProcessingMachine LearningMobile ApplicationMushroom ClassificationConference Object10.1007/978-3-031-86162-8_122-s2.0-105003903693