What is this about? In this post, I walk through a typical workflow using word vectors to analyze unstructured text documents. I also show how a dimensionality reduction technique like TSNE can be leveraged to visualize patterns in the vector space. Why should I care? If you know nothing about... [Read More]
Where the Machines are Looking
When to trust a computer vision model.
What is this about? In this post, I probe the decision-making process of a neural network model, and expose an example of errant model performance in one of the most popular pre-trained networks. Why should I care? Neural networks are often considered to be black-box models, whose decisions are made... [Read More]
Object Detection and Segmentation in Python with Mask-RCNN
Visual examples and discussion on the uses of object detection and object segmentation with Mask R-CNN.
What is this about? In this post, I walk through some hands-on examples of object detection and object segmentation using Mask R-CNN. Why should I care? Besides being super cool, object segmentation can be an incredibly useful tool in a computer vision pipeline. Say you are training a CV model... [Read More]