用可调双正交小波提升CNN的细节捕捉能力
Biorthogonal Tunable Wavelet Unit with Lifting Scheme in Convolutional Neural Network
- 基于提升方案设计灵活小波单元，打破正交与长度限制
- 在CIFAR-10和DTD上分别提升2.12%和9.73%准确率
- 适合需要精细特征提取的图像分类与异常检测任务
本文提出一种基于提升方案构建的新型双正交可调小波单元，放松了正交性和滤波器长度相等的约束，提升了滤波器设计灵活性。该单元增强了卷积、池化和下采样操作，在卷积神经网络中显著改善了图像分类与异常检测性能。当集成至18层残差网络（ResNet-18）时，对CIFAR-10的分类准确率提升2.12%，对可描述纹理数据集（DTD）提升9.73%；在ResNet-34上也获得类似增益。在MVTec异常检测数据集的榛子类别上，该方法在分割与检测任务中均实现优异且平衡的性能，精度与鲁棒性优于现有方法。
原文摘要 · Abstract (English)
This work introduces a novel biorthogonal tunable wavelet unit constructed using a lifting scheme that relaxes both the orthogonality and equal filter length constraints, providing greater flexibility in filter design. The proposed unit enhances convolution, pooling, and downsampling operations, leading to improved image classification and anomaly detection in convolutional neural networks (CNN). When integrated into an 18-layer residual neural network (ResNet-18), the approach improved classification accuracy on CIFAR-10 by 2.12% and on the Describable Textures Dataset (DTD) by 9.73%, demonstrating its effectiveness in capturing fine-grained details. Similar improvements were observed in ResNet-34. For anomaly detection in the hazelnut category of the MVTec Anomaly Detection dataset, the proposed method achieved competitive and wellbalanced performance in both segmentation and detection tasks, outperforming existing approaches in terms of accuracy and robustness.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。