What is a transposed convolution ('deconv') and its checkerboard artifact issue?

hard

Answer

$A convolution that upsamples by inserting zeros between pixels then applying a normal conv — output is larger than input.$
$Used in decoders (U-Net upsampling), GAN generators, and dense prediction.$
$Checkerboard artifacts occur when kernel size isn't divisible by stride: some output pixels get contributions from more kernel positions than others.$
$\mathrm{kernel}_{\mathrm{size}}\; = \;2\; \cdot \;\mathrm{stride}$

Check yourself — multiple choice

Same as pooling
Learnable upsampling — beware checkerboard artifacts when kernel isn't divisible by stride
Reduces spatial size
Has no learnable params

Transposed conv upsamples; kernel/stride mismatch → checkerboard artifacts.

#cnn#generative#segmentation

214 interview questions in this topic.