Although deep convolutional neural networks (DCNNs) have powerful capability of learning complex feature representations, they are limited by poor ability in handling large rotations and scale transformations. In this paper, we propose a novel alternative to conventional convolutional layer named Gabor convolutional layer (GCL) to enhance the robustness to transformations. The GCL is a simple but efficient combination of Gabor prior knowledge and parameters learning. A GCL is composed of three components: Gabor extraction module, weight-sharing convolution module, and transformation pooling module, respectively. DCNNs integrated with GCLs, referred to as transformation-invariant Gabor convolutional networks (TI-GCNs), can be easily built by replacing standard convolutional layers with designed GCLs. Our experimental results on various real-world recognition tasks indicate that encoding traditional hand-crafted Gabor filters with dominant orientation and scale information into DCNNs is of great importance for learning compact feature representations and reinforcing the resistance to scale changes and orientation variations. The source code can be found at https://github.com/GuichenLv .

中文翻译：

---

变换不变 Gabor 卷积网络

尽管深度卷积神经网络 (DCNN) 具有强大的学习复杂特征表示的能力，但它们在处理大旋转和尺度变换方面的能力较差。在本文中，我们提出了一种名为 Gabor 卷积层 (GCL) 的传统卷积层的新替代方案，以增强对变换的鲁棒性。GCL 是 Gabor 先验知识和参数学习的简单而有效的组合。一个 GCL 由三个部分组成：分别是 Gabor 提取模块、权重共享卷积模块和变换池化模块。与 GCL 集成的 DCNN，称为变换不变 Gabor 卷积网络 (TI-GCN)，可以通过用设计的 GCL 替换标准卷积层来轻松构建。我们在各种现实世界识别任务上的实验结果表明，将具有主导方向和尺度信息的传统手工制作的 Gabor 滤波器编码到 DCNN 中对于学习紧凑的特征表示和增强对尺度变化和方向变化的抵抗力非常重要。源代码可以在 https://github.com/GuichenLv 找到。