1 | initial version |
all 3 algorithms are using a simple (1) Nearest Neighbour
classification, like this:
for each image in trainset:
if distance between train & test img is the smallest so far:
keep best id & distance
return best id & distance