The robust generalization capacity of deep Convolutional Neural Networks (CNNs) may compromise the reliability of prediction results derived from knowledge distillation methods. This phenomenon arises from the extensive feature learning capabilities of CNN models during training, which lead to more intricate decision boundaries and diminished disparities between teacher and student models. To address this problem, we put forward a methodology named Feature Regulation-based Reverse Distillation (FRRD). This approach incorporates a Collaborative Disparity Optimization (CDO) module alongside a Selective Feature (SF) component. During the training process, the CDO module ensures consistency by minimizing the feature distance for normal pixels while simultaneously enhancing the discrimination of abnormal pixels through increased feature distances. This strategy facilitates a clear distinction among pixels belonging to different classes. Furthermore, the SF module captures correlations among features across various scales, thereby reducing the influence of anomalous information on subsequent inference processes. As a result, the model’s generalization capacity is augmented and its dependability is elevated. Our FRRD method has been appraised using two publicly accessible datasets, namely MVTecAD and BTAD. When it comes to the MVTecAD dataset, our model secures an AU-ROCPL score of 98.7%. On the BTAD dataset, it reaches an AU-ROCPL score of 97.5%.

Finding more specific subcategories within a larger category is the goal of fine-grained image classification (FGIC), and the key is to find local discriminative regions of visual features. Most existing methods use traditional convolutional operations to achieve fine-grained image classification. However, traditional convolution cannot extract multi-scale features of an image and existing methods are susceptible to interference from image background information. Therefore, to address the above problems, this paper proposes an FGIC model (Attention-PCNN) based on hybrid attention mechanism and pyramidal convolution. The model feeds the multi-scale features extracted by the pyramidal convolutional neural network into two branches capturing global and local information respectively. In particular, a hybrid attention mechanism is added to the branch capturing global information in order to reduce the interference of image background information and make the model pay more attention to the target region with fine-grained features. In addition, the mutual-channel loss (MC-LOSS) is introduced in the local information branch to capture fine-grained features. We evaluated the model on three publicly available datasets CUB-200-2011, Stanford Cars, FGVC-Aircraft, etc. Compared to the state-of-the-art methods, the results show that Attention-PCNN performs better.