Using Channel and Network Layer Pruning Based on Deep Learning for Real-Time Detection of Ginger Images

Fang, Lifa; Wu, Yanqiang; Li, Yuhua; Guo, Hongen; Hua, Zhang; Wang, Xiaoyu; Rui, Xiaoting; Hou, Jialin

doi:10.3390/agriculture11121190

Cited by 14 publications

(12 citation statements)

References 42 publications

(45 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It is easy to achieve a trade-off between performance and parameters by adjusting the pruning rate. At the same time, layer-wise pruning is often used as an auxiliary pruning method in networks with many branches [85], [86]. Although the channel/filter-wise pruning based on reinforcement learning is time-consuming, it effectively solves the limitation of manually set magnitudes [70].…”

Section: Network Pruningmentioning

confidence: 99%

A Review of AIoT-based Edge Devices and Lightweight Deployment

Sun¹,

Wang²,

Zhao³

2023

Preprint

View full text Add to dashboard Cite

<p>Artificial Intelligence of Things (AIoT) which inte?grates artificial intelligence (AI) and the Internet of Things (IoT), has attracted increasing attention recently. With the remarkable development of AI, deep neural networks (DNNs) have achieved great success from research to deployment in many applications. However, deploying complex and state-of-the-art (SOTA) AI models on edge applications is increasingly a big challenge. This paper focuses on the convergence of AIoT-based edge devices, lightweight DNNs, and neural network compression. We provide a comprehensive analysis of them and many practical suggestions for researchers: how to obtain/design lightweight DNNs, how to select suitable AI edge devices, and how to compress and deploy them in practice. Finally, future trends and opportunities are presented, including trustworthy AI, federated learning, and robust deployment. </p>

show abstract

Section: Network Pruningmentioning

confidence: 99%

A Review of AIoT-based Edge Devices and Lightweight Deployment

Sun¹,

Wang²,

Zhao³

2023

Preprint

View full text Add to dashboard Cite

show abstract

“…Channel pruning is also one of the ideas of the model lightweight, to build a lightweight detection model for apple fruitlet pre-thinning, Dandan Wang et al [24] conducted channel pruning based on YOLOv5s, which reduced the number of parameters by 92.7% while maintaining the accuracy of detection. Fang L et al [25] built a lightweight model by pruning the redundant channel and network layers of the YOLOV3 model, and realized real-time monitoring of ginger sprouts and seeds with a map of 98.0%, greatly optimizing the model size and detection time.…”

Section: Model Lightweightmentioning

confidence: 99%

A lightweight detection model for greenhouse-cultivated strawberries based on YOLOv5

Xue

2023

Preprint

View full text Add to dashboard Cite

The efficient detection of strawberries has great significance in the realization of strawberry production estimation and automatic picking in the greenhouse. Factors such as the complex growing environment in the field and fruit aggregation shading, especially for immature strawberries at the turning stage and mature strawberries with high similarity in shape, size, and even color, resulted in low accuracy of detection. This research aims at establishing an improved lightweight model based on YOLOv5 which is for strawberry ripeness detection in natural environments, in response to the problems of some current detection models whose structures are complex and difficult to deploy on low-cost devices. The method first adds a smaller target detection layer to the original network, then removes part of the deep structure of the network to reduce the complexity of the model, by adjusting the depth-multiple and width-multiple partially to achieve a lighter model while ensuring accuracy. After that, the performance of the model was further improved by introducing the BiFPN structure and SimAM attention module. The experimental results show that the improved model provides an effective method for detecting strawberries in natural environments. Compared with the YOLOv5s model, the improved model has a 60.35% reduction in parameters and a 55.47% reduction in the model size, and the mAP, mAP0.5:0.95, and F1 are improved to 91.86%, 79.04%, and 90.9%, respectively. The model proposed in this research proves an excellent generalization performance in three strawberry test sets and provides a reference for subsequent deployment in small/low-cost picking robots.

show abstract

“…The improved model outperformed base YOLOv5 in precision in identifying flower, green, and red tomatoes by 17%, 2%, and 2.3%, respectively, with a decreased model size of 10.5 MB. Fang et al [13] suggested an improved YOLOv3 model to detect ginger shoots and seeds in real-time, which significantly compresses network size and improves inference time by pruning redundant channels and network layers. The results suggest a decrement in model size by 87.2%, an increase in detection speed by 85%, and mAP lagging slightly by 0.1%.…”

Section: Introductionmentioning

confidence: 99%

Lightweight Improved YOLOv5s-CGhostnet for Detection of Strawberry Maturity Levels and Counting

Tamrakar,

Karki,

Kang

et al. 2024

AgriEngineering

View full text Add to dashboard Cite

A lightweight strawberry detection and localization algorithm plays a crucial role in enabling the harvesting robot to effectively harvest strawberries. The YOLO model has often been used in strawberry fruit detection for its high accuracy, speed, and robustness. However, some challenges exist, such as the requirement for large model sizes, high computation operation, and undesirable detection. Therefore, the lightweight improved YOLOv5s-CGhostnet was proposed to enhance strawberry detection. In this study, YOLOv5s underwent comprehensive model compression with Ghost modules GCBS and GC3, replacing modules CBS and C3 in the backbone and neck. Furthermore, the default GIOU bounding box regressor loss function was replaced by SIOU for improved localization. Similarly, CBAM attention modules were added before SPPF and between the up-sampling and down-sampling feature fusion FPN–PAN network in the neck section. The improved model exhibited higher mAP@0.5 of 91.7% with a significant decrement in model size by 85.09% and a reduction in GFLOPS by 88.5% compared to the baseline model of YOLOv5. The model demonstrated an increment in mean average precision, a decrement in model size, and reduced computation overhead compared to the standard lightweight YOLO models.

show abstract

Using Channel and Network Layer Pruning Based on Deep Learning for Real-Time Detection of Ginger Images

Cited by 14 publications

References 42 publications

A Review of AIoT-based Edge Devices and Lightweight Deployment

A Review of AIoT-based Edge Devices and Lightweight Deployment

A lightweight detection model for greenhouse-cultivated strawberries based on YOLOv5

Lightweight Improved YOLOv5s-CGhostnet for Detection of Strawberry Maturity Levels and Counting

Contact Info

Product

Resources

About