Face detection using Viola Jones algorithm (Mar '20)
- Fined tuned pretrained model of VGG16 using TensorFlow and Keras (PyThon, Keras, Tensorflow, Cuda)
- Trained the model using 10,000 real time images obtained from NC State research lab (ARoS)
- Increased the training speed by utilizing the GPU (Nvidia RTX 2070 Max-Q) using CUDA and CuDNN.
- Obtained an accuracy of 86% on the test images.