A Study of YOLO Algorithm for Target Detection

Wen, Haokang; Dai, Fengzhi; Yuan, Yasheng

doi:10.5954/icarob.2021.os13-9

Cited by 20 publications

(12 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, Yolo remains the perfect algorithm when it comes to the combination of both speed as well as efficiency, and accuracy. Yolo is a combination of object detection [27], object localization, and object classification in an image efficiently.…”

Section: The Proposed Methodmentioning

confidence: 99%

Arabic and Latin license plate detection and recognition based on YOLOv7 and image processing methods

Moussaoui

Akkad

Benslimane

2023

Preprint

View full text Add to dashboard Cite

License plate detection has always been challenging, especially with the increasing number of smart radars on roads and highways. By 2021, over 1 billion smart traffic radars have been provided worldwide for the gendarmerie as well as for road control officers from the National Security. This paper introduces a novel Arabic and Latin license plate detection and recognition method. In the first stage of the proposed method and after gathering images to build a new dataset, we used YOLOv7 to detect and localize the license plate in the image. The dataset has been labeled manually before feeding it to the detection system. Second, we applied some of the machine learning algorithms to enhance the detected license plate. For this, we employed thresholding as well as the kernel algorithms in order to remove the additional vertical lines in the plate. Afterward, we used the Arabic OCR as well as Easy OCR techniques to recognize the Arabic and the Latin letters on the license plate. The proposed detection algorithm achieved a precision of 97% and a recall of 98%, consequently, we achieved an F1 score of 98% in license plate detection. For image segmentation, while using Arabic OCR and Easy OCR to segment and extract characters from the detected license plate, we achieved an accuracy of 99 %. The proposed method gave satisfactory results for both detection and segmentation.

show abstract

Section: The Proposed Methodmentioning

confidence: 99%

Arabic and Latin license plate detection and recognition based on YOLOv7 and image processing methods

Moussaoui

Akkad

Benslimane

2023

Preprint

View full text Add to dashboard Cite

show abstract

“…The neck module uses the Path Aggregation Network (PANet) to generate a feature pyramid and bi-directionally fuse low-level spatial features with high-level semantic features to enhance the detection ability of objects at different scales. The head module is responsible for generating target prediction boxes to determine the category, coordinates, and confidence level of the detected object ( Wen et al., 2021 ). Its network contains four network structures of different sizes (YOLOv5s, YOLOv5m, YOLOv5l, and YOLOv5x), thus allowing the user to choose the appropriate model according to their actual needs ( Song et al., 2021 ).…”

Section: Methodsmentioning

confidence: 99%

Recognition of terminal buds of densely-planted Chinese fir seedlings using improved YOLOv5 by integrating attention mechanism

Guo

Wei

et al. 2022

Front. Plant Sci.

View full text Add to dashboard Cite

Accurate and timely information on the number of densely-planted Chinese fir seedlings is essential for their scientific cultivation and intelligent management. However, in the later stage of cultivation, the overlapping of lateral branches among individuals is too severe to identify the entire individual in the UAV image. At the same time, in the high-density planting nursery, the terminal bud of each seedling has a distinctive characteristic of growing upward, which can be used as an identification feature. Still, due to the small size and dense distribution of the terminal buds, the existing recognition algorithm will have a significant error. Therefore, in this study, we proposed a model based on the improved network structure of the latest YOLOv5 algorithm for identifying the terminal bud of Chinese fir seedlings. Firstly, the micro-scale prediction head was added to the original prediction head to enhance the model’s ability to perceive small-sized terminal buds. Secondly, a multi-attention mechanism module composed of Convolutional Block Attention Module (CBAM) and Efficient Channel Attention (ECA) was integrated into the neck of the network to enhance further the model’s ability to focus on key target objects in complex backgrounds. Finally, the methods including data augmentation, Test Time Augmentation (TTA) and Weighted Boxes Fusion (WBF) were used to improve the robustness and generalization of the model for the identification of terminal buds in different growth states. The results showed that, compared with the standard version of YOLOv5, the recognition accuracy of the improved YOLOv5 was significantly increased, with a precision of 95.55%, a recall of 95.84%, an F1-Score of 96.54%, and an mAP of 94.63%. Under the same experimental conditions, compared with other current mainstream algorithms (YOLOv3, Faster R-CNN, and PP-YOLO), the average precision and F1-Score of the improved YOLOv5 also increased by 9.51-28.19 percentage points and 15.92-32.94 percentage points, respectively. Overall, The improved YOLOv5 algorithm integrated with the attention network can accurately identify the terminal buds of densely-planted Chinese fir seedlings in UAV images and provide technical support for large-scale and automated counting and precision cultivation of Chinese fir seedlings.

show abstract

“…While, YOLOv3 has made two improvements on its basis YOLOv2. One is to continue to use the residual model to deepen the network structure; the other is to use FPN Architecture for multi-scale detection [19].…”

Section: You Only Look Once (Yolo)mentioning

confidence: 99%

“…Coronavirus disease (COVID- 19) was discovered in December 2019 and was first reported by the Wuhan municipal government in China [1]. The transmission of the COVID-19 virus has a very fast transmission speed, because the virus can be transmitted by droplets [2].…”

Section: Introductionmentioning

confidence: 99%

Mask Detection on Motorcyclists Using YOLOv4

Firdauz

Rachmawati

Sulistiyo

2023

bits

View full text Add to dashboard Cite

The use of mask is a mandatory for everyone in the pandemic regulation to prevent the spread of COVID-19 infection. This becomes a pandemic regulation for everyone, especially in public places like in traffic situation, such as pedestrian and motorcyclists. However, many motorcyclists ignore this rule or do not use the mask properly, let alone they have high risk in being infected by the virus; Thus, a computer vision-based solution is required to help monitoring it. This study aims to built a system to automatically detect the use of mask on motorcyclists. Here, we propose a YOLOv4 model, one of YOLO variants, which is popular in the object detection task and featured with a considerably high speed in real-time situation. This study also implements domain adaptation to discuss the object detection performances. Based on the experimental results in various scenarios, our model obtained average accuracy of 78.3% and IoU of 64.8% for class with_mask, average accuracy of 78.4% and IoU of 56.3% for class without_mask, and average accuracy of 87% and IoU of 55.5% for class incorrect_mask

show abstract

A Study of YOLO Algorithm for Target Detection

Cited by 20 publications

References 0 publications

Arabic and Latin license plate detection and recognition based on YOLOv7 and image processing methods

Arabic and Latin license plate detection and recognition based on YOLOv7 and image processing methods

Recognition of terminal buds of densely-planted Chinese fir seedlings using improved YOLOv5 by integrating attention mechanism

Mask Detection on Motorcyclists Using YOLOv4

Contact Info

Product

Resources

About