Artificial intelligence for the selection and classification of fingerlings

The article is the result of a literature review related to artificial vision methods that allows the selection and classification process of the fry, using different important elements and criteria that allow the cameras and software to carry out the selection process. The problem depends on the question: How does artificial vision strengthen the fry selection process? and the objective is to identify the appropriate artificial vision method for the fry selection process. The methodology begins with the analysis of the systematic review that allows us to know the different types of classification methods, with algorithms used in artificial vision and determine the most appropriate one during the fry selection process. The representative result was the transformer model, being the best solution as a complement to artificial vision for the selection of fry with the use of mathematical techniques that influence the classification of the appropriate fry. In conclusion, the Transformer model is considered fundamental during the fry selection process by allowing precision, efficiency, robustness, scalability and flexibility in artificial vision.