<![CDATA[Stunting is a global health problem, especially in developing countries including Indonesia. One of the main causes of stunting is malnutrition, especially in children aged 0-23 months. Therefore, this study aims to develop an AI-based model to detect calories in Indonesian food images for stunting prevention, using the Transfer Learning method with AlexNet. In this article, we propose a new deep learning-based food image calorie detection model called, Alexnet Interactive Transfer Learning (AITL). AITL is built based on Alexnet's Convolution Neural Network architecture, and further modified at the last Convolution layer and classification layer. AITL was evaluated using a dataset from the Indonesian food database. Experiments were conducted on the dataset to detect food types and their calorific content. There are ten classes of authentic Indonesian food types, which include: Rendang, Bika Ambon, Pempek, Sate Ayam, Gado-gado, Ayam Pop, Kerak Telor, Rawon, Lemang, and Ayam Betutu. The accuracy of the developed AITL model reached 95.33%. The results of the tests conducted show that Alexnet-based AITL outperforms other CNNs in terms of accuracy and efficiency.]]>
