Stride length estimation is one of the most crucial aspects of Pedestrian Dead Reckoning (PDR). Due to the measurement noise of inertial sensors, individual variances of pedestrians, and the uncertainty in pedestrians walking, there is a substantial error in the assessment of stride length, which causes the accumulated deviation of Pedestrian Dead Reckoning (PDR). With the help of multi-gait analysis, which decomposes strides in time and space with greater detail and accuracy, a novel and revolutionary stride estimating model or scheme could improve the performance of PDR on different users. This paper presents a diverse stride gait dataset by using inertial sensors that collect foot movement data from people of different genders, heights, and walking speeds. The dataset contains 4690 walking strides data and 19,083 gait labels. Based on the dataset, we propose a threshold-independent stride segmentation algorithm called SDATW and achieve an F-measure of 0.835. We also provide the detailed results of recognizing four gaits under different walking speeds, demonstrating the utility of our dataset for helping train stride segmentation algorithms and gait detection algorithms.

Currently, supervised person re-identification (Re-ID) models trained on labeled datasets can achieve high recognition performance in the same data domain. However, accuracy drops dramatically when these models are directly applied to other unlabeled datasets or natural environments, due to a significant sample distribution gap between the two domains. Unsupervised Domain Adaptation (UDA) methods can solve this problem by fine-tuning the model on the target dataset with pseudo-labels generated by the clustering method. Yet, these methods are primarily aimed at the image-based person Re-ID domain. This is because the background noise and interference information are complex and changeable in the video scenarios, resulting in large intra-class distances and small inter-class spaces, which easily lead to noisy labels. Huge domain gap and noisy labels hinder clustering and training processes heavily in the video-based person Re-ID. To address the problem, we propose a novel UDA method via Dynamic Clustering and Co-segment Attentive Learning (DCCAL) for it. DCCAL includes a Dynamic Clustering (DC) module and a Co-segment Attentive Learning (CAL) module to alleviate noisy labels by clustering pedestrians adaptively within different generation processes and reducing domain gap with a co-segmentation-based attention mechanism, respectively. Additionally, we introduce Kullback-Leibler (KL) divergence loss to reduce the distribution of features between two domains for better performance. Experimental results on two large-scale video-based person Re-ID datasets, MARS and DukeMTMC-VideoReID (DukeV), demonstrate exceptional precision performance. Our method outperforms state-of-the-art semi-supervised and unsupervised approaches by 1.1% in Rank-1 and 1.5% in mAP on DukeV, as well as 3.1% and 2.1% in Rank-1 and mAP on MARS, respectively.

Current attention or transform modules in Convolutional Neural Networks (CNNs) are designed pursuing lightweight and in-place. Generally, we need to decrease the channel dimension of input feature maps for reducing computation cost firstly. And then we do some transformation for extracting weight maps or converting to other feature space etc. Finally, we increase the channel dimension back for outputting feature maps with the same size as input. When we change the channel dimension, commonly we choose 1×1 convolutional layers or fully connected layers. They are simple and effective, but need learning parameters and consuming more memory with other computation resources. We propose a novel parameter free method named Channel Transformer Network (CTN) to decrease or increase channels for these modules whilst keeping most information with lower computation complexity. We also introduce a Video Co-segment Attentive Network (VCAN) for person re-identification (ReID) to improve pedestrian’s noticeable representation across multiple video frames. We embed CTN in Non-local, CBAM, COSAM and VCAN blocks to replace 1×1 convolutional or fully connected layers. Experiments of VCAN and CTN embedding models on Mars dataset for person ReID show significant performance in computation efficiency and accuracy, especially VCAN reaches 90.05% in Rank-1. We believe CTN can also be used in other vision tasks like image classification and object detection etc.

Few-shot object detection (FSOD) is proposed to solve the application problem of traditional detectors in scenarios lacking training samples. The meta-learning methods have attracted the researchers’ attention for their excellent generalization performance. They usually select the same class of support features according to the query labels to weight the query features. However, the model cannot possess the ability of active identification only by using the same category support features, and feature selection causes difficulties in the testing process without labels. The single-scale feature of the model also leads to poor performance in small object detection. In addition, the hard samples in the support branch impact the backbone’s representation of the support features, thus impacting the feature weighting process. To overcome these problems, we propose a multi-scale feature fusion and attentive learning (MSFFAL) framework for few-shot object detection. We first design the backbone with multi-scale feature fusion and channel attention mechanism to improve the model’s detection accuracy on small objects and the representation of hard support samples. Based on this, we propose an attention loss to replace the feature weighting module. The loss allows the model to consistently represent the objects of the same category in the two branches and realizes the active recognition of the model. The model no longer depends on query labels to select features when testing, optimizing the model testing process. The experiments show that MSFFAL outperforms the state-of-the-art (SOTA) by 0.7–7.8% on the Pascal VOC and exhibits 1.61 times the result of the baseline model in MS COCO’s small objects detection.

The emergence of Transformer has led to the rapid development of video understanding, but it also brings the problem of high computational complexity. Previously, there were methods to divide the feature maps into windows along the spatiotemporal dimensions and then calculate the attention. There are also methods to perform down-sampling during attention computation to reduce the spatiotemporal resolution of features. Although the complexity is effectively reduced, there is still room for further optimization. Thus, we present the Windows and Linear Transformer (WLiT) for efficient video action recognition, by combining Spatial-Windows attention with Linear attention. We first divide the feature maps into multiple windows along the spatial dimensions and calculate the attention separately inside the windows. Therefore, our model further reduces the computational complexity compared with previous methods. However, the perceptual field of Spatial-Windows attention is small, and global spatiotemporal information cannot be obtained. To address this problem, we then calculate Linear attention along the channel dimension so that the model can capture complete spatiotemporal information. Our method achieves better recognition accuracy with less computational complexity through this mechanism. We conduct extensive experiments on four public datasets, namely Something-Something V2 (SSV2), Kinetics400 (K400), UCF101, and HMDB51. On the SSV2 dataset, our method reduces the computational complexity by 28% and improves the recognition accuracy by 1.6% compared to the State-Of-The-Art (SOTA) method. On the K400 and two other datasets, our method achieves SOTA-level accuracy while reducing the complexity by about 49%.

The Non-local self-attention mechanism can significantly improve the capability of feature representation with long-range dependencies at the cost of high computational complexity. To address the issue, the self-attention-based autoregressive axial transformer has been proposed to apply attention along a single axis of the feature maps instead of the whole ones with large receptive fields. It performs axial-attention twice along the height- and width-axis respectively in the spatial dimension of the feature maps for the image data. However, there is still room for improvement. We can convert the 2D spatial feature map into a 1D feature sequence and just perform axial-attention once along it to save more computing resources. Motivated by the insight, we propose an Efficient Axial-Attention Network (EAAN) for video-based person re-identification (Re-ID) to reduce computation and improve accuracy by serializing feature maps with multi-granularity and reducing the number of axial-attention runs. We also introduce a deserialization approach that can restore the shape of the feature maps. Moreover, we expand the CTN (Channel Transformer Network) to a wider range of uses. Additionally, we verify that the serialized feature sequence can enhance expressiveness in our EAAN with lower complexity. Experiments on MARS and DukeMTMC-VideoReID (DukeV) datasets show outstanding performance in computation efficiency and accuracy. It not only outperforms the state-of-the-art method on MARS by 0.3% in both Rank-1 and mAP, and surpasses that on DukeV by 0.1% in Rank-1 with equal mAP, but also reduces parameters and GFLOPS (Giga Floating-point Operations Per Second) by 16.9% and 6.6% respectively compared to another axial-attention-based method. The code will be available at https://github.com/hopstone/EAAN.