search

arunponnusamy / gender-detection-keras

Gender detection (from scratch) using deep learning with keras and cvlib.
http://arunponnusamy.com
MIT License
184 stars 68 forks source link

Feature Request #6

Open ekinburak opened 6 years ago

ekinburak commented 6 years ago

Hey Arun,

Great Lib ! Can you please share more info about "training our own models" ? How can i do that ?

Thank you in advance,

arunponnusamy commented 6 years ago

Hello @ekinburak , have a look at the training section in readme. You can swap in your own dataset for training. If you want to increase the number of classes you can change that in train.py script.

Hyperclaw79 commented 6 years ago

@arunponnusamy That information is helpful for people already experienced in ML. But for total beginners like me, a more comprehesive guide like @ekinburak suggested would definitely be useful.

ekinburak commented 6 years ago

@arunponnusamy , alright i will take a look on these files. The reason I asked this is I believe we need to improve train set. (I am having wrong predictions, but very low percentage.) Have you used data augmentation models on your train set ?

GreatAbbas commented 5 years ago

Thanks for your nice code, its really great and does a really good job for me as a student. But I am really new to python and image processing and I don't know at all what can I do to improve the dataset because women use a special type of dress for their clothing in our region and it covers a part of their face, also they are rarely tagged as woman and I think its because of our people face difference (maybe the people you showed to it were different ...) can you give me a really NOOB-FRIENDLY guide for how to add data to dataset ( or maybe train it if they are the same) ???

Thanks

arunponnusamy commented 5 years ago

Have a look at this post. Might be helpful. https://www.learnopencv.com/age-gender-classification-using-opencv-deep-learning-c-python/