We provide a high bandwidth, alias-free convolutional kernel parameterization with learnable kernel size and constant parameter cost.
We present the first edition of "VIPriors: Visual Inductive Priors for Data-Efficient Deep Learning" challenges. We offer four data-impaired challenges to encourage data efficient solutions.
We add line priors through a trainable Hough transform block into a deep network.
We show we can prevent CNNs from exploiting absolute location through image boundary effects.