Enhancing geometric modeling in convolutional neural networks: limit deformable convolution
Open Access
- 4 March 2025
- journal article
- research article
- Published by Springer Nature in Complex & Intelligent Systems
- Vol. 11 (4) , 1-14
Abstract
Convolutional neural networks (CNNs) are constrained in their capacity to model geometric transformations due to their fixed geometric structure. To overcome this problem, researchers introduce deformable convolution, which allows the convolution kernel to be deformable on the feature map. However, deformable convolution may introduce irrelevant contextual information during the learning process and thus affect the model performance. DCNv2 introduces a modulation mechanism to control the diffusion of the sampling points to control the degree of contribution of offsets through weights, but we find that such problems still exist in practical use. Therefore, we propose a new limit deformable convolution to address this problem, which enhances the model ability by adding adaptive limiting units to constrain the offsets and adjusts the weight constraints on the offsets to enhance the image-focusing ability. In the subsequent work, we perform lightweight work on the limit deformable convolution and design three kinds of LDBottleneck to adapt to different scenarios. The limit deformable network, equipped with the optimal LDBottleneck, demonstrated an improvement in mAP75 of 1.4% compared to DCNv1 and 1.1% compared to DCNv2 on the VOC2012+2007 dataset. Furthermore, on the CoCo2017 dataset, different backbones equipped with our limit deformable module achieved satisfactory results. The source code for this work is publicly available at https://github.com/1977245719/LDCN.Keywords
Funding Information
- Shenyang Young and Middle-aged Science and Technology Innovation Talent Support Program (RC220452)
This publication has 25 references indexed in Scilit:
- Deformable ConvNets V2: More Deformable, Better Results2019
- ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design2018
- CBAM: Convolutional Block Attention Module2018
- MobileNetV2: Inverted Residuals and Linear Bottlenecks2018
- ShuffleNet: An Extremely Efficient Convolutional Neural Network for Mobile Devices2018
- Xception: Deep Learning with Depthwise Separable Convolutions2017
- Deep Residual Learning for Image Recognition2016
- Rethinking the Inception Architecture for Computer Vision2016
- Going deeper with convolutions2015
- Microsoft COCO: Common Objects in Context2014