AI-Driven mobile application for enhancing efficiency and preserving herbal knowledge

Traditional Thai medicine and local herbal knowledge are important local wisdom heritages, yet there are still limitations in access and methods of knowledge transfer. Therefore, this research aims to develop a mobile application for herbal information search by photographs. This application collects 30 ancient herbal medicine recipes from 4 provinces in the lower northeastern region, namely Surin, Buriram, Chaiyaphum, and Nakhon Ratchasima. The application is developed by using Extreme Programming approach which applies Convolutional neural network (CNN: DenseNet201) model on TensorFlow. The model is trained with 4,211 herbal images covering 30 species, with an accuracy of 92%. The system includes a function that allows users to add new herbal data, which must be validated by 5 users. In addition, there is a learning media on herbal medicine recipes in multimedia format developed according to the ADDIE Model process. The test results from a sample group of 100 people who passed the purposive sampling criteria, who must have had experience using herbal medicine at least 5 times and have basic knowledge of information technology. The evaluation results found that there was the highest level of satisfaction in all 3 areas: usability and accessibility (=4.25, SD=0.66), system efficiency (=4.32, SD=0.71), and content quality (=4.36, SD=0.71). The results of relationship analysis between basic factors and satisfaction levels revealed that the service area was the only factor that had a statistically significant effect on overall satisfaction (p < 0.05). The results of the research demonstrated that the application of artificial intelligence in combination with multimedia learning media in the application was possible to increase the perception of herbal information. This study demonstrates creative science by integrating traditional herbal knowledge with modern mobile application technology, resulting in a practical and culturally relevant tool developed in the Thai language. GRAPHICAL ABSTRACT HIGHLIGHTS Innovative Integration of AI and Herbal Knowledge, The study introduces an AI-driven mobile application using Convolutional Neural Networks (CNN). Multimedia-Based Collaborative Learning Platform, The application combines image and video infographics based on traditional Thai herbal wisdom. Community-Centered Data Contribution System, The system allows users to verify and add new herbal data. Strong Evaluation Results Across Multiple Dimensions, Satisfaction surveys from 100 participants across four provinces indicated high ratings in usability, performance, and content quality (Mean > 4.25 in all aspects), reflecting the app’s effectiveness and relevance.