Very simple model. I retrained the top layer of the Inception v3 model neural network in order to tell if the person in a given picture was a man or a woman. Tensorflow provides already trained NNs with a retraining python program which allows you to retrain the network to custom classes. The ultimate goal of this project is to integrate this network into a useful program. FULLY CUSTOM MODEL/TRAINING PROGRAM TO COME
- Install Tensorflow
- Place picture you wish to classify in root of repo, or alter path in point 3 (--image=$HOME/) to desired directory
- Run following command from tensorflow directory
python tensorflow/examples/label_image/label_image.py \
--graph=/tmp/output_graph.pb --labels=/tmp/output_labels.txt \
--input_layer=Mul \
--output_layer=final_result \
--input_mean=128 --input_std=128 \
--image=$HOME/**<your photo>**
- Class probabilities should output to console.
Images were gathered using a batch image chrome extension. Keywords were simply "Man" and "Woman" so the classifier is representative of Google Image search results, not my opinion.