The main challenge for fine-grained few-shot image classification is to learn feature representations with higher inter-class and lower intra-class variations, with a mere few labelled samples. Conventional few-shot learning methods however cannot be naively adopted for this fine-grained setting -- a quick pilot study reveals that they in fact push for the opposite (i.e., lower inter-class variations and higher intra-class variations). To alleviate this problem, prior works predominately use a support set to reconstruct the query image and then utilize metric learning to determine its category. Upon careful inspection, we further reveal that such unidirectional reconstruction methods only help to increase inter-class variations and are not effective in tackling intra-class variations. In this paper, we for the first time introduce a bi-reconstruction mechanism that can simultaneously accommodate for inter-class and intra-class variations. In addition to using the support set to reconstruct the query set for increasing inter-class variations, we further use the query set to reconstruct the support set for reducing intra-class variations. This design effectively helps the model to explore more subtle and discriminative features which is key for the fine-grained problem in hand. Furthermore, we also construct a self-reconstruction module to work alongside the bi-directional module to make the features even more discriminative. Experimental results on three widely used fine-grained image classification datasets consistently show considerable improvements compared with other methods. Codes are available at: https://github.com/PRIS-CV/Bi-FRN.

Deep learning methods have been widely used in the field of fault diagnosis of rotating machinery. As one of the deep learning methods, multi-scale convolutional neural network (MSCNN) can achieve accurate fault diagnosis by extracting feature information of vibration signals at different scales. However, in practical applications, a small number of fault samples limit the performance of MSCNN. To overcome this problem, a small sample rotating machinery fault diagnosis method based on multi-branch and multi-scale dynamic convolutional network (MBSDCN) is proposed. Firstly, a feature splitting strategy is proposed, which splits a single input into multi-branch inputs to ensure that each multi-scale convolutional layer is matched to the appropriate input. Secondly, a channel reconstruction (CR) attention mechanism is designed, and CR attention mechanism can form a direct mapping of high-dimensional space with the fully connected layer through the reconstruction processing of channels, thus making the connection between feature channels closer. Finally, a novel multi-scale feature extraction model is constructed, in which multi-scale convolutional layers are applied to extract multi-branch input features, and the CR attention mechanism is used to calibrate the contribution of multi-scale convolution layers. The performance of MBSDCN is evaluated with CWRU rolling bearing dataset and UOC gearbox dataset and is compared with some advanced deep learning methods. The comparison results indicate that the proposed MBSDCN can accurately diagnose faults under small sample conditions. In addition, the experimental results in noisy environments indicate that the proposed MBSDCN has satisfactory anti-noise performance under small sample conditions and noisy environments.

Fault diagnosis of rotating machinery is difficult under the strong noisy environment. Although singular value decomposition (SVD) can remove noise from vibration signals, the singular value threshold is normally determined by expert experience. To solve this problem, a fault diagnosis method based on average descent rate (ADR)-SVD and two-dimensional residual neural network (Resnet) is proposed. First, ADR-SVD uses the ADR index to construct the singular value descent rate difference spectrum and uses the maximum value of the spectrum as the singular value threshold. The noise reduction process of ADR-SVD requires little expert experience. Then, in order to adaptively identify the fault features of the signals, we introduce Gramian angular difference field (GADF), which can transform the one-dimensional signals into two-dimensional images and preserve the temporal correlation of the one-dimensional signals. Finally, we construct a two-dimensional Resnet to learn image features and identify fault types. The proposed method is tested on Case Western Reserve University (CWRU) bearing dataset, Driveline Dynamic Simulator (DDS) gearbox dataset, and University of Connecticut (UoC) gearbox dataset under the strong noisy environment, which achieves the accuracies of 98.00%, 99.00%, and 98.88%, respectively. The accuracies of other deep learning methods and singular value difference spectrum method are below 95%. The comparisons show that the proposed method has better noise reduction effect and can diagnose the fault type more accurately.

The main challenge for fine-grained few-shot image classification is to learn feature representations with higher inter-class and lower intra-class variations, with a mere few labelled samples. Conventional few-shot learning methods however cannot be naively adopted for this fine-grained setting – a quick pilot study reveals that they in fact push for the opposite (i.e., lower inter-class variations and higher intra-class variations). To alleviate this problem, prior works predominately use a support set to reconstruct the query image and then utilize metric learning to determine its category. Upon careful inspection, we further reveal that such unidirectional reconstruction methods only help to increase inter-class variations and are not effective in tackling intra-class variations. In this paper, we introduce a bi-reconstruction mechanism that can simultaneously accommodate for inter-class and intra-class variations. In addition to using the support set to reconstruct the query set for increasing inter-class variations, we further use the query set to reconstruct the support set for reducing intra-class variations. This design effectively helps the model to explore more subtle and discriminative features which is key for the fine-grained problem in hand. Furthermore, we also construct a self-reconstruction module to work alongside the bi-directional module to make the features even more discriminative. We introduce the snapshot ensemble method in the episodic learning strategy – a simple trick to further improve model performance without increasing training costs. Experimental results on three widely used fine-grained image classification datasets, as well as general and cross-domain few-shot image datasets, consistently show considerable improvements compared with other methods. Codes are available at https://github.com/PRIS-CV/BiEN .

Remaining useful life (RUL) prediction is crucial in the field of engineering, which can reduce the frequency of accidents and the maintenance cost of machinery. With the increasing complexity of rotating machinery, the data analysis methods based on deep learning have become the mainstream methods of prediction work. However, most of the current RUL prediction methods only use single-sensor data as input, which cannot effectively use multi-sensor data. In addition, as an advanced deep learning prediction method, temporal convolutional network (TCN) only uses the past time information of vibration data to determine the current health status of bearings, while ignoring the importance of future time information of vibration data. To solve the above problems, a bearing RUL prediction method based on multi-sensor data fusion and bidirectional-temporal attention convolutional network (Bi-TACN) is proposed in this paper. In multi-sensor data fusion, multi-sensor data are combined into multi-channel data, and a channel-weighted attention is designed to emphasize the importance of each sensor data. Compared with traditional multi-sensor data fusion, the proposed fusion method allows deep prediction networks to learn more useful feature information from multi-sensor data. Then, Bi-TACN is developed to predict the RUL of bearings. Bi-TACN is mainly composed of the forward TCN block and the backward TCN block, both of which can learn the past and future time information of multi-sensor data simultaneously. Moreover, a temporal attention mechanism is embedded in Bi-TACN to adaptively calibrate the weights of the two TCN blocks, so as to achieve dynamic feature fusion of past and future time information. RUL prediction experiments are carried out through Xi’an Jiao tong University bearing dataset and PHM 2012 bearing dataset respectively. Compared with the advanced prediction methods, the proposed method can accurately predict the RUL of more types of bearings and has low prediction errors.