Having extracted all images, it’s interesting to see the variation in
color, lighting, etc. of the material.
A downsampled subset of example images that the convolutional neural net
will train a classifier on. Personal illustration by author, November 2018.
Twitter Facebook Google+ LinkedIn
My initial working assumption was that I could exclude the part of finding
the exact location of the goal/canvas in each frame as
it is of secondary impor...
In most cases there is an obvious difference between the canvas before
having being hit by a puck and after. The question, raised in a
previous post,
is how ...
Yesterday, I recorded seven rounds in artificial light. This is yet another
tricky situation I wanted to capture on video in order to later be able to
test f...