Real-time detection network for tiny traffic sign using multi-scale attention module

Yang, Tingting; Tong, Chao

doi:10.1007/s11431-021-1950-9

Cited by 24 publications

(10 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The preprocessing module is an essential part of traffic sign recognition methods [19,21]. The image resolution obtained by the in-vehicle camera is generally huge.…”

Section: The Feature-enhanced Preprocessing Modulementioning

confidence: 99%

“…a brown‐toned or upset‐down traffic sign) when the feature space of the augmented dataset does not overlap with that of the test set. For the challenging small size of traffic signs, methods such as [21] use a cropping strategy to reduce the resolution of the input images for the training and testing process (e.g. one 2048x2048 image can be cropped into four 512x512 images).…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A feature‐enhanced hybrid attention network for traffic sign recognition in real scenes

He,

Lan,

Zhou

et al. 2024

IET Image Processing

View full text Add to dashboard Cite

Currently, traffic sign recognition techniques have been brought into the assistive driving of automobiles. However, small traffic sign recognition in real scenes is still a challenging task due to the class imbalance issue and the size limit of the traffic signs. To address the above issues, a feature‐enhanced hybrid attention network is proposed based on YOLOv5s for a small, fast, and accurate traffic sign detector. First, a series of online data augmentation strategies are designed in the preprocessing module for the model training. Second, the hybrid channel and spatial attention module CSAM are integrated into the backbone for a better feature extraction ability. Third, the channel attention module CAM is used in the detection head for a more efficient feature fusion ability. To validate the approach, extensive experiments are conducted based on the Tsinghua‐Tencent 100K dataset. It is found that the novel method achieves state‐of‐the‐art performance with only negligible increases in the model parameter and computational overhead. Specifically, the , parameters, and FLOPs are 85.8%, 7.13 M, and 16.1 G, respectively.

show abstract

“…The preprocessing module is an essential part of traffic sign recognition methods [19,21]. The image resolution obtained by the in-vehicle camera is generally huge.…”

Section: The Feature-enhanced Preprocessing Modulementioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

A feature‐enhanced hybrid attention network for traffic sign recognition in real scenes

He,

Lan,

Zhou

et al. 2024

IET Image Processing

View full text Add to dashboard Cite

show abstract

“…The main characteristic of the attention module is that it can help to strengthen the feature extraction ability of a CNN model, and thus enhance the model's analysis and classification performance. Since it is a computer-aided mechanism inspired by the human beings' visualization system [31], the attention module can easily recognize which regions are significant and carry important information in an image, just like the ability of the human eye. Then, it will help the deep learning model to concentrate on learning the important information only by applying focus on these regions of interest (ROIs), such as assigning a higher degree of importance to them.…”

Section: Attention Modulementioning

confidence: 99%

Double-Enhanced Convolutional Neural Network for Multi-Stage Classification of Alzheimer’s Disease

Wong,

Abdullah,

Shapiai

2024

Adv. sci. technol. eng. syst. j.

View full text Add to dashboard Cite

Being known as an irreversible neurodegenerative disease which has no cure to date, detection and classification of Alzheimer's disease (AD) in its early stages is significant so that the deterioration process can be slowed down. Generally, AD can be classified into three major stages, ranging from the "normal control" stage with no symptoms shown, the "mild cognitive impairment (MCI)" stage with minor symptoms, and the AD stage which depicts major and serious symptoms. Due to its generative features, MCI patients tend to easily progress to the AD stage if appropriate diagnosis and prevention measures are not taken. However, it is difficult to accurately identify and diagnose the MCI stage due to its mild and insignificant symptoms that often lead to misdiagnosis. In other words, the classification of multiple stages of AD has been a challenge for medical professionals. Thus, deep learning models like convolutional neural networks (CNN) have been popularly utilized to overcome this challenge. Nevertheless, they are still limited by the issue of limited medical images and their weak feature representation ability. In this study, a doubleenhanced CNN model is proposed by incorporating an attention module and a generative adversarial network (GAN) to classify magnetic resonance imaging (MRI) brain images into 3 classes of AD. MRI images are obtained from the Open Access Series of Imaging Studies (OASIS) database and four experiments are done in this study to observe the classification performance of the enhanced model. From the results obtained, it can be observed that the enhanced CNN model with GAN and attention module has achieved the best performance of 99% as compared to the other models. Hence, this study has shown that the double-enhanced CNN model has effectively boosted the performance of the deep learning model and overcame the challenge in the multi-stage classification of AD.

show abstract

“…At present, the above two methods are commonly used for traffic sign detection. Yang et al 23 designed a visual multi-scale attention module for traffic sign detection, which integrates multi-scale feature maps with channel weights and spatial masks. Zhang et al 24 designed a bottom-up enhancement path to enhance the feature pyramid, thereby effectively utilizing fine-grained features at the bottom to achieve precise positioning of traffic signs.…”

Section: Related Workmentioning

confidence: 99%

Improved YOLOv5-based for small traffic sign detection under complex weather

Qu,

Yang,

Zhou

et al. 2023

Sci Rep

View full text Add to dashboard Cite

Traffic sign detection is a challenging task for unmanned driving systems. In the traffic sign detection process, the object size and weather conditions vary widely, which will have a certain impact on the detection accuracy. In order to solve the problem of balanced detecting precision of traffic sign recognition model in different weather conditions, and it is difficult to detect occluded objects and small objects, this paper proposes a small object detection algorithm based on improved YOLOv5s in complex weather. First, we add the coordinate attention(CA) mechanism in the backbone, a light-weight yet effective module, embedding the location information of traffic signs into the channel attention to improve the feature extraction ability of the network. Second, we exploit effectively fine-grained features about small traffic signs from the shallower layers by adding one prediction head to YOLOv5s. Finally, we use Alpha-IoU to improve the original positioning loss CIoU, improving the accuracy of bbox regression. Applying this model to the recently proposed CCTSDB 2021 dataset, for small objects, the precision is 88.1%, and the recall rate is 79.8%, compared with the original YOLOv5s model, it is improved by 12.5% and 23.9% respectively, and small traffic signs can be effectively detected under different weather conditions, with low miss rate and high detection accuracy. The source code will be made publicly available at https://github.com/yang-0706/ImprovedYOLOv5s.

show abstract

Real-time detection network for tiny traffic sign using multi-scale attention module

Cited by 24 publications

References 40 publications

A feature‐enhanced hybrid attention network for traffic sign recognition in real scenes

A feature‐enhanced hybrid attention network for traffic sign recognition in real scenes

Double-Enhanced Convolutional Neural Network for Multi-Stage Classification of Alzheimer’s Disease

Improved YOLOv5-based for small traffic sign detection under complex weather

Contact Info

Product

Resources

About