| layout | title | og_rel_url | og_rel_image | og_description |
|---|---|---|---|---|
default |
SeeForTwo: Computer Vision |
assets/img/SeeForTwo_eyes_640x320_margin.jpg |
Detecting left and right eyes is an experiment in an area of personal interest, understanding the minimal requirements for network (Convolutional Neural Network, CNN / Deep Neural Network, DNN) structures for particular object detection (computer vision) applications. |
Detecting left and right eyes describes an experiment in an area of personal interest, understanding the minimal requirements for network (Convolutional Neural Network, CNN / Deep Neural Network, DNN) structures for particular object detection (computer vision) applications. A spin-out of this work is https://github.com/SeeForTwo/hier_object, a Python library for putting annotations in a hierarchy that groups ground truth object bounding boxes. The library is well tested using property based testing, random testing for hierarchical annotations.
The title "SeeForTwo" comes for the informal standard of judging computer vision performance using a comparison against a human that only has two seconds to look at the image. For example, https://cloud.google.com/vision/automl/docs/prepare says:
AutoML Vision models can't generally predict labels that humans can't assign. So, if a human can't be trained to assign labels by looking at the image for 1-2 seconds, the model likely can't be trained to do it either.
I can be contacted via a gmail account that matches my github account. Any other use of "SeeForTwo" on the internet is not related.