Video object detection (VOD) focuses on detecting objects for each frame in a video, which is a challenging task due to appearance deterioration in certain video frames. Recent works usually distill crucial information from multiple support frames to improve the reference features, but they only perform at frame level or proposal level that cannot integrate spatial-temporal features sufficiently. To deal with this challenge, we treat VOD as a spatial-temporal hierarchical features interacting process and introduce a Multi-level Spatial-Temporal (MST) feature aggregation framework to fully exploit frame-level, proposal-level, and instance-level information in a unified framework. Specifically, MST first measures context similarity in pixel space to enhance all frame-level features rather than only update reference features. The proposal-level feature aggregation then models object relation to augment reference object proposals. Furthermore, to filter out irrelevant information from other classes and backgrounds, we introduce an instance ID constraint to boost instance-level features by leveraging support object proposal features that belong to the same object. Besides, we propose a Deformable Feature Alignment (DAlign) module before MST to achieve a more accurate pixel-level spatial alignment for better feature aggregation. Extensive experiments are conducted on ImageNet VID and UAVDT datasets that demonstrate the superiority of our method over state-of-the-art (SOTA) methods. Our method achieves 83.3% and 62.1% with ResNet-101 on two datasets, outperforming SOTA MEGA by 0.4% and 2.7%.

Filter pruning for a pre-trained convolutional neural network is most normally performed through human-made constraints or criteria such as norms, ranks, etc. Typically, the pruning pipeline comprises two-stage: first learn a sparse structure from the original model, then optimize the weights in the new prune model. One disadvantage of using human-made criteria to prune filters is that the design and selection of threshold criteria depend on complicated prior knowledge. Besides, the pruning process is less robust due to the impact of directly regularizing on filters. To address the problems mentioned, we propose an effective one-stage pruning framework: introducing a trainable collaborative layer to jointly prune and learn neural networks in one go. In our framework, we first add a binary collaborative layer for each original filter. Then, a new type of gradient estimator - asymptotic gradient estimator is first introduced to pass the gradient in the binary collaborative layer. Finally, we simultaneously learn the sparse structure and optimize the weights from the original model in the training process. Our evaluation results on typical benchmarks, CIFAR and ImageNet, demonstrate very promising results against other state-of-the-art filter pruning methods.

Durability and reliability are the major bottlenecks of the proton-exchange-membrane fuel cell (PEMFC) for large-scale commercial deployment. With the help of prognostic approaches, we can reduce its maintenance cost and maximize its lifetime. This paper proposes a hybrid prognostic method for PEMFCs based on a decomposition forecasting framework. Firstly, the original voltage data is decomposed into the calendar aging part and the reversible aging part based on locally weighted regression (LOESS). Then, we apply an adaptive extended Kalman filter (AEKF) and long short-term memory (LSTM) neural network to predict those two components, respectively. Three-dimensional aging factors are introduced in the physical aging model to capture the overall aging trend better. We utilize the automatic machine-learning method based on the genetic algorithm to train the LSTM model more efficiently and improve prediction accuracy. The aging voltage is derived from the sum of the two predicted voltage components, and we can further realize the remaining useful life estimation. Experimental results show that the proposed hybrid prognostic method can realize an accurate long-term voltage-degradation prediction and outperform the single model-based method or data-based method.