The main objective of this study is to develop and deploy an automated process of categorizing bananas according to their level of maturity or ripeness using computer vision techniques. This is done using advanced computer vision techniques with keen interest in exploring different types of transfer learning applied to pre-trained models. There is a lot of literature written about fruit-sorting systems using different technologies. This study focuses on the latest and most advanced machine learning concepts, and computer vision techniques which will produce a system able to perform efficiently in real-time. In this study banana dataset that is readily available online is used. The dataset is fed into a pre-trained model and results are studied. Different types of transfer learning are used to perform comparative analysis. Optimization techniques will be deployed to hyperparameters to foster high level accuracy performance. After model optimization, the model will be trained and then tested using the validation dataset. This paper provides an overview of existing literature, and results on how advanced computer vision techniques, coupled with transfer learning techniques, can be used to achieve high accuracy levels on pre-trained models.