A　gender　classification　system　uses　a　given　image　from　human　face　to　tell　the　gender　of　the　given　person.　An　effective　gender　classification　approach　is　able　to　improve　the　performance　of　many　other　applications,　including　image　or　video　retrieval,　security　monitoring,　human-computer　interaction　and　so　on.　In　this　paper,　an　effective　method　for　gender　classification　task　in　frontal　facial　images　based　on　convolutional　neural　networks　(CNNs)　is　presented.　Our　experiments　have　been　shown　that　the　method　of　CNNs　for　gender　classification　task　is　effective　and　achieves　higher　classification　accuracy　than　others　on　FERET　and　CAS-PEAL-R1　facial　datasets.　Finally,　we　built　a　gender　classification　demo,　where　input　is　the　scene　image　per　frame　captured　by　the　camera　and　the　output　is　the　original　scene　image　with　marked　on　detected　facial　areas.　©　Springer　International　Publishing　Switzerland　2015.