December 2018
Intermediate to advanced
764 pages
18h 18m
English
Dilated convolutions need bigger input and hence, are memory intensive. This presents computational problems when using high-resolution pictures. Reid et al. (https://arxiv.org/pdf/1611.06612.pdf) prosed a method called RefiNet to overcome this problem which is shown below:

RefiNet uses an encoder followed by a decoder. Encoder outputs of CNN. The decoder concatenates the features of various sizes:

The concatenation is done upscaling the low dimensional ...
Read now
Unlock full access