Deep Learning as a Solution for Egg Quality Assessment

The society is aware of the essential nutrients present in eggs, such as protein, vitamins, and minerals. The eggshell can indicate and be utilized to detect and assess egg quality. Dirty eggs pose a significant risk to consumers as they can harbor harmful Salmonella bacteria, leading to adverse health consequences. Automating the process of selecting and classifying clean and dirty eggs is an alternative solution that is more efficient, as manual labor is time-consuming, complex, and monotonous. Therefore, it is necessary to develop an automated egg classification system. This research aims to develop a classification process for dirty and clean eggs. The methodology employs GLCM-Deep Learning to automatically identify dirty and clean eggs. Three optimizers are used: Adam optimizer, rmsprop optimizer, and sgd optimizer, with the RELU activation function. The dataset consists of a total of 1000 images of dirty eggs and 1000 clean eggs, with 400 samples in the test dataset. The best result from the GLCM-Deep Learning model was achieved using the rmsprop optimizer with an accuracy of 62.5%. This research is of great importance as it addresses the complexity and inefficiency in egg classification and selection processes, and enables automatic determination and evaluation of egg quality.