Learning a Discriminative Filter Bank Within a CNN for Fine-Grained Recognition

Abstract

Compared to earlier multistage frameworks using CNN features, recent end-to-end deep approaches for finegrained recognition essentially enhance the mid-level learning capability of CNNs. Previous approaches achieve this by introducing an auxiliary network to infuse localization information into the main classification network, or a sophisticated feature encoding method to capture higher order feature statistics. We show that mid-level representation learning can be enhanced within the CNN framework, by learning a bank of convolutional filters that capture class-specific discriminative patches without extra part or bounding box annotations. Such a filter bank is well structured, properly initialized and discriminatively learned through a novel asymmetric multi-stream architecture with convolutional filter supervision and a non-random layer initialization. Experimental results show that our approach achieves state-of-the-art on three publicly available fine-grained recognition datasets (CUB-200-2011, Stanford Cars and FGVC-Aircraft). Ablation studies and visualizations are provided to understand our approach.

On the Impact of Object and Sub-Component Level Segmentation Strategies for Supervised Anomaly Detection within X-Ray Security Imagery

“…From the results presented in Tables I and II, we can observe that a sub-component level segmentation strategy, supported by the secondary fine-grain CNN classification of DFL model [22], offers significantly superior anomaly detection performance (A: 97.91, TP: 98.20, FP: 3.50 - Table I) than an object level segmentation strategy overall (Table II). Furthermore, fine-grain CNN classification similarly offers the highest overall accuracy and lowest false positive rate (A: 89.77, FP: 3.88 - Table II) for object level segmentation.…”

“…Second stage binary classification via CNN performed less well overall with the sub-component segmentation strategy (lower accuracy (A) caused by significantly higher false positive (FP) - Table II). Fine grain classification model (DFL [22]) offer the lowest false positive and maximal accuracy for both segmentation strategies (Table I). We can deduce that increased levels isolation via segmentation to the sub-component level improves the performance of the discriminative feature space learnt by the fine-grain technique [20]- [22] whilst more classical object classification CNN architectures perform only marginally better on objects than sub-components (Table II).…”

“…Fine grain classification model (DFL [22]) offer the lowest false positive and maximal accuracy for both segmentation strategies (Table I). We can deduce that increased levels isolation via segmentation to the sub-component level improves the performance of the discriminative feature space learnt by the fine-grain technique [20]- [22] whilst more classical object classification CNN architectures perform only marginally better on objects than sub-components (Table II). Figure 4 illustrates the attention (red/pink patch with the highest focus) of the fine-grain DFL model [22] whilst trained on object-level and sub-component level segmentation data respectively.…”

The Devil is in the Channels: Mutual-Channel Loss for Fine-Grained Image Classification

“…Recently, some frameworks employ a more general architecture that can localize discriminative parts within an image without any extra supervision from part annotations, and thus it makes the fine-grained image classification more feasible in real-world scenarios. Wang et al [40] claimed that improving mid-level convolutional feature representation can bring significant advantages for part-based fine-grained classification. This is accomplished by introducing a bank of discriminative filters in the classical convolutional neural networks (CNNs) architecture and it can be trained in an endto-end fashion.…”

“…We can observe that feature channels after going through the MC-Loss become class-aligned and each focues on different discriminative regions that roughly correspond to object parts. subtle differences is the key for solving fine-grained image classification [40], [46], [47].…”