posted an update

We were also able to use Scikit Learn to analyze the pixels of faces returned by the Bing (source) + HP Haven (facial) API.

We prepared a record manually using our web interface for labeling faces, recorded with a binary 1 or 0 indicating distress. After doing so, tried out Naive Bayes, SVM, and KNN classifiers for detecting distressed faces with our manual human labels. While our sample size of images was small, this approach easily scales with more images or real time drone video feeds, something allowed with more time.

Log in or sign up for Devpost to join the conversation.