Irregular Convolutional Neural Networks

Abstract

Convolutional kernels are basic and vital components of deep Convolutional Neural Networks (CNN). In this paper, we equip convolutional kernels with shape attributes to generate the deep Irregular Convolutional Neural Networks (ICNN). Compared to traditional CNN applying regular convolutional kernels like $3\times 3$, our approach trains irregular kernel shapes to better fit the geometric variations of input features. In other words, shapes are learnable parameters in addition to weights. The kernel shapes and weights are learned simultaneously during end-to-end training with the standard back-propagation algorithm. Experiments for semantic segmentation are implemented to validate the effectiveness of our proposed ICNN.

7 pages, 5 figures, 3 tables

References in corpus (8)

• Deep Residual Learning for Image Recognition
• Going Deeper with Convolutions
• The Cityscapes Dataset for Semantic Urban Scene Understanding
• Fully Convolutional Networks for Semantic Segmentation
• Understanding the Effective Receptive Field in Deep Convolutional Neural Networks
• Rethinking the Inception Architecture for Computer Vision
• Deformable Convolutional Networks
• Accelerating Deep Convolutional Networks using low-precision and sparsity

Cited by in corpus (1)

• H-CNN: Spatial Hashing Based CNN for 3D Shape Analysis