This paper proposes the Parallel Residual Bi-Fusion Feature Pyramid Network (PRB-FPN) for fast and accurate single-shot object detection. Feature Pyramid (FP) is widely used in recent visual detection, however the top-down pathway of FP cannot preserve accurate localization due to pooling shifting. The advantage of FP is weakened as deeper backbones with more layers are used. In addition, it cannot keep up accurate detection of both small and large objects at the same time. To address these issues, we propose a new parallel FP structure with bi-directional (top-down and bottom-up) fusion and associated improvements to retain high-quality features for accurate localization. We provide the following design improvements: 1) parallel bifusion FP structure with a bottom-up fusion module (BFM) to detect both small and large objects at once with high accuracy; 2) concatenation and re-organization (CORE) module provides a bottom-up pathway for feature fusion, which leads to the bi-directional fusion FP that can recover lost information from lower-layer feature maps; 3) CORE feature is further purified to retain richer contextual information. Such CORE purification in both top-down and bottom-up pathways can be finished in only a few iterations; 4) adding of a residual design to CORE leads to a new Re-CORE module that enables easy training and integration with a wide range of deeper or lighter backbones. The proposed network achieves state-of-the-art performance on the UAVDT17 and MS COCO datasets.

Despite recent progress in Multiple Object Tracking (MOT), several obstacles such as occlusions, similar objects, and complex scenes remain an open challenge. Meanwhile, a systematic study of the cost-performance tradeoff for the popular tracking-by-detection paradigm is still lacking. This paper introduces SMILEtrack, an innovative object tracker that effectively addresses these challenges by integrating an efficient object detector with a Siamese network-based Similarity Learning Module (SLM). The technical contributions of SMILETrack are twofold. First, we propose an SLM that calculates the appearance similarity between two objects, overcoming the limitations of feature descriptors in Separate Detection and Embedding (SDE) models. The SLM incorporates a Patch Self-Attention (PSA) block inspired by the vision Transformer, which generates reliable features for accurate similarity matching. Second, we develop a Similarity Matching Cascade (SMC) module with a novel GATE function for robust object matching across consecutive video frames, further enhancing MOT performance. Together, these innovations help SMILETrack achieve an improved trade-off between the cost (e.g., running speed) and performance (e.g., tracking accuracy) over several existing state-of-the-art benchmarks, including the popular BYTETrack method. SMILETrack outperforms BYTETrack by 0.4-0.8 MOTA and 2.1-2.2 HOTA points on MOT17 and MOT20 datasets. Code is available at http://github.com/pingyang1117/SMILEtrack_official.

Change detection (CD) aims to find the difference between two images at different times and output a change map to represent whether the region has changed or not. To achieve a better result in generating the change map, many State-of-The-Art (SoTA) methods design a deep learning model that has a powerful discriminative ability. However, these methods still get lower performance because they ignore spatial information and scaling changes between objects, giving rise to blurry boundaries. In addition to these, they also neglect the interactive information of two different images. To alleviate these problems, we propose our network, the Scale and Relation-Aware Siamese Network (SARAS-Net) to deal with this issue. In this paper, three modules are proposed that include relation-aware, scale-aware, and cross-transformer to tackle the problem of scene change detection more effectively. To verify our model, we tested three public datasets, including LEVIR-CD, WHU-CD, and DSFIN, and obtained SoTA accuracy. Our code is available at https://github.com/f64051041/SARAS-Net.

Wearing a safety helmet is important in construction and manufacturing industrial activities to avoid unpleasant situations. This safety compliance can be ensured by developing an automatic helmet detection system using various computer vision and deep learning approaches. Developing a deep-learning-based helmet detection model usually requires an enormous amount of training data. However, there are very few public safety helmet datasets available in the literature, in which most of them are not entirely labeled, and the labeled one contains fewer classes. This paper presents the Safety HELmet dataset with 5K images (SHEL5K) dataset, an enhanced version of the SHD dataset. The proposed dataset consists of six completely labeled classes (helmet, head, head with helmet, person with helmet, person without helmet, and face). The proposed dataset was tested on multiple state-of-the-art object detection models, i.e., YOLOv3 (YOLOv3, YOLOv3-tiny, and YOLOv3-SPP), YOLOv4 (YOLOv4 and YOLOv4_pacsp-x-mish), YOLOv5-P5 (YOLOv5s, YOLOv5m, and YOLOv5x), the Faster Region-based Convolutional Neural Network (Faster-RCNN) with the Inception V2 architecture, and YOLOR. The experimental results from the various models on the proposed dataset were compared and showed improvement in the mean Average Precision (mAP). The SHEL5K dataset had an advantage over other safety helmet datasets as it contains fewer images with better labels and more classes, making helmet detection more accurate.

Neural networks have enabled state-of-the-art approaches to achieve incredible results on computer vision tasks such as object detection. However, such success greatly relies on costly computation resources, which hinders people with cheap devices from appreciating the advanced technology. In this paper, we propose Cross Stage Partial Network (CSPNet) to mitigate the problem that previous works require heavy inference computations from the network architecture perspective. We attribute the problem to the duplicate gradient information within network optimization. The proposed networks respect the variability of the gradients by integrating feature maps from the beginning and the end of a network stage, which, in our experiments, reduces computations by 20% with equivalent or even superior accuracy on the ImageNet dataset, and significantly outperforms state-of-the-art approaches in terms of AP50 on the MS COCO object detection dataset. The CSPNet is easy to implement and general enough to cope with architectures based on ResNet, ResNeXt, and DenseNet. Source code is at https://github.com/WongKinYiu/CrossStagePartialNetworks.

The recent growth of the elderly population has led to the requirement for constant home monitoring as solitary living becomes popular. This protects older people who live alone from unwanted instances such as falling or deterioration caused by some diseases. However, although wearable devices and camera-based systems can provide relatively precise information about human motion, they invade the privacy of the elderly. One way to detect the abnormal behavior of elderly residents under the condition of maintaining privacy is to equip the resident’s house with an Internet of Things system based on a non-invasive binary motion sensor array. We propose to concatenate external features (previous activity and begin time-stamp) along with extracted features with a bi-directional long short-term memory (Bi-LSTM) neural network to recognize the activities of daily living with a higher accuracy. The concatenated features are classified by a fully connected neural network (FCNN). The proposed model was evaluated on open dataset from the Center for Advanced Studies in Adaptive Systems (CASAS) at Washington State University. The experimental results show that the proposed method outperformed state-of-the-art models with a margin of more than 6.25% of the F₁ score on the same dataset.