Anemia is a disease present worldwide. High prevalence of anemia (43%) is found in the child population and its main long-term effect (slow cognitive development) can remain even if the disease has disappeared. One of the main reasons for the high prevalence of anemia in Peru is the poor screening coverage during the growth of the child due to the parents' fear of infringing pain on their children. We take advantage that anemia produces pallor in the hands, fingers and ungueal bed to develop a semaphore for this disease. This screening tool uses photographic images of the patient's ungueal bed to determine if they have a high, medium or low possibility of having anemia. Sixty people participated in the study and 6 photographic images for each participant's right hand were captured. The images were processed to extract regions of interest from each of the fingernails. Datasets were generated and a neural network was used to predict the risk of anemia. Initial results show that the proposed semaphore of anemia reaches a sensitivity of 0.79 and specificity of 0.91. These results indicate that the semaphore of anemia may be used as a screening method to reduce the number of blood tests and the time of evaluation from 15 minutes (rapid test with portable hemoglobinometer) to 1 minute.