y everyone 4-6? do i need to change some parameter?

[kelisima001]

[brmarkus]

Excuse me, can you rephrase your question, please? Can you describe what you see, what you have expected instead? Which application do you use, how have you called it, which parameter have you used, have you changed something, do you use the same NN-model underneath?

[kelisima001]

i used python3.8 opencv 4.5.3.56 i downloaded age_net.caffemodel and gender_net.caffemodel and then putted under AgeGender else did not change anything then i used python AgeGender.py --input "C:\Users\xxx\Pictures\2110021F20422A.jpg" show this

[kelisima001]

i git clone the project and cd to AgeGender in windows

[brmarkus]

Cannot comment on the age-network's precision or what material was used for training, whether it used "representative" images, including Asian faces of women in different ages... Can you try other images showing different men, woman of different age, please?

Random examples: https://upload.wikimedia.org/wikipedia/commons/2/2d/Angela_Merkel_Juli_2010_-_3zu4.jpg https://upload.wikimedia.org/wikipedia/commons/a/af/Claudia_Schiffer_C%C3%A9sars_1993.jpg

[chinlee1523]

I found the same thing

[brmarkus]

You might contact "https://github.com/GilLevi/AgeGenderDeepLearning" and check "https://talhassner.github.io/home/publication/2015_CVPR" for which material was used to train the model. It could mean that the selected data is not representative enough.