When we say "combinations of features" and "each possible subset" what do we mean? Would it be possible to explain the last sentence more?
motoole2
In the context of identifying lines in images, the "features" represent edges detected in images (through a combination of filtering and thresholding). To fit line models to an image, we could try to identify the subset of features (edges) corresponding to a particular line and fitting the line model to this subset. Unfortunately, this would be extremely expensive, because it requires that we check all combinations of edges in order to do so. So, as stated in this slide, "it is not feasible to check all combinations of features (edges) by fitting a (line) model to each possible subset".
When we say "combinations of features" and "each possible subset" what do we mean? Would it be possible to explain the last sentence more?
In the context of identifying lines in images, the "features" represent edges detected in images (through a combination of filtering and thresholding). To fit line models to an image, we could try to identify the subset of features (edges) corresponding to a particular line and fitting the line model to this subset. Unfortunately, this would be extremely expensive, because it requires that we check all combinations of edges in order to do so. So, as stated in this slide, "it is not feasible to check all combinations of features (edges) by fitting a (line) model to each possible subset".