Academic and scientific research activities have a growing economic importance, requiring larger outlays and investments both in advanced and emerging nations. In both cases it is a matter of national prestige but more than that, of strategic relevance, since the presence of large numbers of highly educated citizens contributing to the advance of human knowledge has well-established impacts on technological and industrial capabilities. These, in turn, are highly relevant to ensure the competitiveness and the economic security of nations as well as yielding other benefits to the economy (Dasgupta and David (1994) and Salter and Martin (2001)). The care and promotion of research activities should thus be a focus of public policies aimed at ensuring social and economic development (Stephan (2012) and Etzkowitz (2013)). A particularly pressing issue in this matter is to assess the quality of the production generated by researchers in all fields of knowledge. On one hand, it is of interest to detect areas in which nationals have international impact, as to concentrate resources on them.

When you have trained your detection model, how well do you know about the learning of your model. The best way to valid your model is to predict the unseen pictures and get a performance to measure it. Usually we have three main indexes, Precision, Recall and F1. They are so much important to measure your model. When you have trained your detection model, how well do you know about the learning of your model.