Stop-band Energy Constraint for Orthogonal Tunable Wavelet Units in Convolutional Neural Networks for Computer Vision problems
This work addresses image classification and anomaly detection for computer vision applications, particularly on texture-rich datasets, with incremental improvements over existing methods.
The paper tackles the problem of improving image classification and anomaly detection in CNNs for texture-rich datasets by introducing a stop-band energy constraint for filters in orthogonal tunable wavelet units, resulting in accuracy gains of 2.48% on CIFAR-10 and 13.56% on the Describable Textures dataset.
This work introduces a stop-band energy constraint for filters in orthogonal tunable wavelet units with a lattice structure, aimed at improving image classification and anomaly detection in CNNs, especially on texture-rich datasets. Integrated into ResNet-18, the method enhances convolution, pooling, and downsampling operations, yielding accuracy gains of 2.48% on CIFAR-10 and 13.56% on the Describable Textures dataset. Similar improvements are observed in ResNet-34. On the MVTec hazelnut anomaly detection task, the proposed method achieves competitive results in both segmentation and detection, outperforming existing approaches.