相关参考文献
注意：仅列出部分参考文献，下载原文获取全部文献信息。
Article
Computer Science, Information Systems
Shengwang Wu et al.
Summary: With increasing government support, the VR/AR industry has developed rapidly. Gesture recognition, an important human-computer interaction method in this industry, faces challenges of low recognition accuracy and speed. To address these issues, an improved YOLOv5s algorithm is proposed, utilizing content-aware re-assembly of features (CARAFE) and adaptive spatial feature fusion (ASFF) to improve accuracy and efficiency. The introduced bottleneck transformer method reduces parameters while increasing accuracy and acceleration. The improved algorithm achieves 96.8% mAP and higher confidence in detection results compared to the original YOLOv5s algorithm.
Article
Agronomy
Efrem Yohannes Obsie et al.
Summary: Early detection and accurate rating of plant diseases are crucial for protecting crop quality and yield. Traditional methods for identifying mummy berry disease rely on field surveys by experts and growers. However, deep learning models show promise but lack sufficient labeled data for accurate estimation. Additionally, complex visual characteristics and spatial variation pose challenges for disease severity estimation.
Article
Chemistry, Multidisciplinary
Zhuang Li et al.
Summary: In this paper, an improved two-stage industrial defect detection framework is proposed, which completes positioning and classification tasks through two specific models. Multiple sets of comparative experiments are conducted on various datasets, and the results show significant improvement. The framework achieves 83.3% mean average precision (mAP) on the Enriched-NEU-DET dataset and 91.0% accuracy in the industrial defect environment.
APPLIED SCIENCES-BASEL
(2022)
Article
Agronomy
Shilin Li et al.
Summary: This paper proposes a lightweight target identification algorithm based on improved YOLOv5, which improves the accuracy of the model by improving data enhancement methods and introducing attention mechanisms. The algorithm reduces model complexity by using Deep Separable Convolution. Experimental results show that the algorithm outperforms other common target detection algorithms in terms of average precision and model size.
Article
Computer Science, Artificial Intelligence
Yanyun Shen et al.
Summary: This study proposes a MIOU loss function to improve the performance of bounding box regression. By considering the Manhattan distance and normalization coefficient, it addresses the shortcomings of existing loss functions. Experimental results in remote sensing and natural object detection scenarios demonstrate the excellent performance of the MIOU loss.
Article
Mathematics
Xingshuai Yang et al.
Summary: The random dumping of garbage in rivers has led to water quality deterioration and environmental impact. This paper proposes the YOLOv5_CBS model for garbage detection in rivers, which improves upon the YOLOv5 model through the addition of a coordinate attention mechanism, a bidirectional feature pyramid network, and a faster loss function. Experimental results demonstrate that the proposed model outperforms other state-of-the-art models in terms of recall, average precision, and F1 score for garbage detection in rivers.
Article
Mathematical & Computational Biology
Hongbin Jiao
Summary: This paper conducts intelligent research on the deep learning recognition method based on the vehicle-road collaborative information interaction system. Through simulation comparison experiments, it demonstrates that deep learning algorithms have higher accuracy and shorter warning time compared to traditional algorithms in vehicle target detection, vehicle flow identification, and emergency decision-making.
COMPUTATIONAL INTELLIGENCE AND NEUROSCIENCE
(2022)
Article
Agronomy
Haiqing Wang et al.
Summary: In this study, a plant disease detection and classification method based on the optimized lightweight YOLOv5 model was proposed to improve the accuracy and efficiency of disease classification. The optimized model showed better performance in terms of accuracy and operation time compared to other mainstream models, demonstrating its effectiveness in plant disease recognition.
Article
Computer Science, Artificial Intelligence
Zhong Qu et al.
Summary: An improved one-stage object detector M-FCFN based on YOLOv5 method is proposed in this paper, achieving significant detection accuracy improvement by cross-layer fusion of multi-scale features for objects of different sizes.
IMAGE AND VISION COMPUTING
(2022)
Article
Chemistry, Analytical
Tasweer Ahmad et al.
Summary: Human action recognition and detection from drones is a challenging task due to image acquisition and processing constraints. We propose a method that combines object recognition and classifier techniques for action identification. Our method outperforms previous approaches on the Okutama dataset, indicating its suitability for this specific dataset.
Article
Chemistry, Multidisciplinary
Iftikhar Ahmad et al.
Summary: Insect pests have a significant impact on agricultural production, and there is a growing need for insect pest detection and classification methods. This study proposes an automatic system based on YOLO object detection for detecting and identifying insect pests in crops using digital images/videos, aiming to reduce reliance on pesticides.
APPLIED SCIENCES-BASEL
(2022)
Article
Agronomy
Zejin Sun et al.
Summary: This paper proposes a tapping trajectory detection method for natural rubber trees based on an improved YOLOv5 model, which includes adding the coordinate attention mechanism, designing the convolution and GhostBottleneck module, and introducing the EIoU loss function. The experimental results show that the improved YOLOv5-CCE model outperforms the original YOLOv5 and other lightweight detection algorithms in terms of overall performance.
Article
Computer Science, Artificial Intelligence
Zanobya N. Khan et al.
Summary: Deep neural networks, such as CNNs, have significantly improved human activity recognition but face challenges of limited labeled samples and high computational costs. A proposed attention-based multi-head model enhances activity recognition accuracy and computational efficiency.
APPLIED SOFT COMPUTING
(2021)
Article
Environmental Sciences
Jianqing Zhao et al.
Summary: This study proposed an improved YOLOv5-based method to accurately detect wheat spikes in UAV images, addressing issues with detection accuracy. The results showed that the proposed method outperforms existing object detection algorithms in improving wheat spike detection accuracy.
Article
Engineering, Marine
Junchi Zhou et al.
Summary: This study improved the YOLOv5s algorithm to enhance the detection accuracy and recall rate of ship images, further realizing the recognition of multiple types of ships, laying the foundation for path planning and automatic obstacle avoidance of unmanned ships.
JOURNAL OF MARINE SCIENCE AND ENGINEERING
(2021)
