Abstract
This study aims to detect the freshness of tilapia using Convolutional Neural Networks (CNN) based on digital images. The process begins by collecting images of tilapia in various freshness conditions, then continues with data preprocessing to ensure optimal data quality for model training. The CNN model is developed and trained using processed images, and then its performance is evaluated using metrics such as accuracy, precision, recall, and F1-score. The results showed that CNN was able to detect visual changes in the fish's eyes, such as color and texture, so that it could distinguish fresh and non-fresh fish with good accuracy. This study shows the potential of CNN in improving the efficiency and accuracy of tilapia freshness detection, which can be applied in the fishery industry to improve the quality and safety of fish products.