In some psychological disorders such as autism and schizophrenia, loss of facial expression recognition skill may complicate patient's daily life. Information technology may help to develop facial expression recognition skill by educational software and games. Our aim was to prepare a reliable human facial expressions digital photograph set, and define the characteristics of people who are successful on recognizing human facial expressions on computer monitor. We have taken 1001 photographs of 40 models, resembling facial expressions of neutral, angry, feared, happy, surprised, disgusted, and sad. By using a web based survey, 427 volunteers have evaluated the photographs. Of 427 users, 275 (64.4%) were female, 152 (35.6%) were male. Their age was 35.5 9.3 (mean standard deviation). They have received a mean consensus score as mean of their consensus scores for each photograph. At the end, we have obtained 356 photographs of facial expressions. The evaluators whose consensus scores below 0.65 were evaluated as poor evaluators and their evaluation about facial expressions were neglected. To understand characteristics of a good evaluator, a logistic regression was applied using input parameters as geographical region, size of settlement, gender, educational level and age. Gender and age were statistically significant factors. Females were more successful in recognition of facial expressions. There was a negative relation with age, younger users were more successful. When this set will be used, one should remind that facial expression recognition performance decrease by age and male gender is a disadvantage for recognition of facial expressions. © 2011 IEEE.