Integrating Multiple Receptive Fields through Grouped Active Convolution

2018/11/11 by Yunho Jeon, Jeon, Yunho, Junmo Kim +1
Computer Science · Engineering · #CCD and CMOS Imaging Sensors #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Neural Networks and Reservoir Computing #Sparse and Compressive Sensing Techniques

paper · pdf · doi:10.48550/arxiv.1811.04387

openalex publication_date 2018/11/11 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Convolutional networks have achieved great success in various vision tasks. This is mainly due to a considerable amount of research on network structure. In this study, instead of focusing on architectures, we focused on the convolution unit itself. The existing convolution unit has a fixed shape and is limited to observing restricted receptive fields. In earlier work, we proposed the active convolution unit (ACU), which can freely define its shape and learn by itself. In this paper, we provide a detailed analysis of the previously proposed unit and show that it is an efficient representation of a sparse weight convolution. Furthermore, we extend an ACU to a grouped ACU, which can observe multiple receptive fields in one layer. We found that the performance of a naive grouped convolution is degraded by increasing the number of groups; however, the proposed unit retains the accuracy even though the number of parameters decreases. Based on this result, we suggest a depthwise ACU, and various experiments have shown that our unit is efficient and can replace the existing convolutions.

Citations

• ImageNet Large Scale Visual Recognition Challenge
• Fast and Accurate Deep Network Learning by Exponential Linear Units (ELUs)
• Wide Residual Networks
• Neural Architecture Search with Reinforcement Learning
• Oriented Response Networks
• MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications
• Rethinking Atrous Convolution for Semantic Image Segmentation
• Caffe: Convolutional Architecture for Fast Feature Embedding
• Learning Multiple Layers of Features from Tiny Images

Related

• Active Convolution: Learning the Shape of Convolution for Image Classification
• Understanding the Effective Receptive Field in Deep Convolutional Neural Networks
• Receptive Field Refinement for Convolutional Neural Networks Reliably Improves Predictive Performance
• RecConv: Efficient Recursive Convolutions for Multi-Frequency Representations
• Large Receptive Field Networks for High-Scale Image Super-Resolution