An Approach for Weed Detection Using CNNs And Transfer Learning

Ofori, Martinson; El‐Gayar, Omar F.

doi:10.24251/hicss.2021.109

Cited by 8 publications

(6 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the first stage of the proposed approach, stateof-the-art models with pre-trained weights were trained on the plant seedling dataset. The result of this baseline (average 91.5% prediction accuracy) was comparable to the earlier result of prior studies that employed the same dataset [42], [51]- [53]. Further, the literature points to a relationship between model accuracy and model compression such that pruning models could result in decreased accuracy [38], [39].…”

Section: Discussionsupporting

confidence: 77%

“…Each techniquetransfer learning; model compression; and ensemble learningdelivers benefits that can enhance and underpin the generalizability of DL to precision agriculture systems. Transfer learning, where a model trained on one task can be ported to another task, offers opportunities for reducing overfitting and ensuring robust results in the face of limited training data [51]. Model compression offers additional benefits for reducing the size of the DL models, which means an equivalent reduction in both energy consumption and inference time [33].…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

A Deep Learning Model Compression and Ensemble Approach for Weed Detection

Ofori¹,

El‐Gayar²,

O'Brien³

et al. 2022

Proceedings of the Annual Hawaii International Conference on System Sciences

Self Cite

View full text Add to dashboard Cite

Site-specific weed management is an important practice in precision agriculture. Current advances in artificial intelligence have resulted in the use of large deep convolutional neural networks for weed detection. In this paper, a transfer learning, model compression, and ensemble learning approach is introduced that is suitable for resource-limited hardware such as mobile and embedded devices. The resulting ensemble model achieves 91.2% classification accuracy which is comparable to the performance of state-of-the-art deep learning models (such as the vanilla VGG16, DenseNet, and ResNet) while being about 62.22% smaller in size than DenseNet (the smallest-sized full-sized model). The approach used in this study is beneficial for further development of deep convolutional neural networks on smaller resource-limited hardware typically used in agriculture, as well as other industries such as healthcare and telecommunication.

show abstract

Section: Discussionsupporting

confidence: 77%

Section: Discussionmentioning

confidence: 99%

A Deep Learning Model Compression and Ensemble Approach for Weed Detection

Ofori¹,

El‐Gayar²,

O'Brien³

et al. 2022

Proceedings of the Annual Hawaii International Conference on System Sciences

Self Cite

View full text Add to dashboard Cite

show abstract

“…To make the algorithms more manageable for hardware with low resources while still retaining accuracy, in this study [ 127 ] the authors used ensemble learning approaches, transfer learning, and model compression. The suggested method was carried out in three steps: transfer learning, pruning-based compression, quantization, and Huffman encoding, and model ensembling with a weighted average for improved accuracy.…”

Section: Summary Of Identified Articles In Slrmentioning

confidence: 99%

Weed Detection Using Deep Learning: A Systematic Literature Review

Murad

Mahmood

Forkan

et al. 2023

Sensors

View full text Add to dashboard Cite

Weeds are one of the most harmful agricultural pests that have a significant impact on crops. Weeds are responsible for higher production costs due to crop waste and have a significant impact on the global agricultural economy. The importance of this problem has promoted the research community in exploring the use of technology to support farmers in the early detection of weeds. Artificial intelligence (AI) driven image analysis for weed detection and, in particular, machine learning (ML) and deep learning (DL) using images from crop fields have been widely used in the literature for detecting various types of weeds that grow alongside crops. In this paper, we present a systematic literature review (SLR) on current state-of-the-art DL techniques for weed detection. Our SLR identified a rapid growth in research related to weed detection using DL since 2015 and filtered 52 application papers and 8 survey papers for further analysis. The pooled results from these papers yielded 34 unique weed types detection, 16 image processing techniques, and 11 DL algorithms with 19 different variants of CNNs. Moreover, we include a literature survey on popular vanilla ML techniques (e.g., SVM, random forest) that have been widely used prior to the dominance of DL. Our study presents a detailed thematic analysis of ML/DL algorithms used for detecting the weed/crop and provides a unique contribution to the analysis and assessment of the performance of these ML/DL techniques. Our study also details the use of crops associated with weeds, such as sugar beet, which was one of the most commonly used crops in most papers for detecting various types of weeds. It also discusses the modality where RGB was most frequently used. Crop images were frequently captured using robots, drones, and cell phones. It also discusses algorithm accuracy, such as how SVM outperformed all machine learning algorithms in many cases, with the highest accuracy of 99 percent, and how CNN with its variants also performed well with the highest accuracy of 99 percent, with only VGGNet providing the lowest accuracy of 84 percent. Finally, the study will serve as a starting point for researchers who wish to undertake further research in this area.

show abstract

“…Regarding greener DL models in applications domains, we can find advances for greener DL-based solutions as well. For instance, for weed detection, Ofori et al combined the mobile-sized EfficientNet with transfer learning to achieve up to 95.44% classification accuracy on plant seedlings [20], and model compression achieving 62.22% smaller in size than DenseNet (the smallest-sized full-sized model) [21]. Moreover, for pig posture classification, Witte et al reported the YOLOv5 model achieving an accuracy of 99,4% for pig detection, and EfficientNet achieving a precision of 93% for pig posture classification [22].…”

Section: Green Aimentioning

confidence: 99%

Energy Efficiency of Training Neural Network Architectures: An Empirical Study

Xu,

Martínez-Fernández,

Martinez

et al. 2023

Proceedings of the Annual Hawaii International Conference on System Sciences

View full text Add to dashboard Cite

The evaluation of Deep Learning (DL) models has traditionally focused on criteria such as accuracy, F1 score, and related measures. The increasing availability of high computational power environments allows the creation of deeper and more complex models. However, the computations needed to train such models entail a large carbon footprint. In this work, we study the relations between DL model architectures and their environmental impact in terms of energy consumed and CO 2 emissions produced during training by means of an empirical study using Deep Convolutional Neural Networks. Concretely, we study: (i) the impact of the architecture and the location where the computations are hosted on the energy consumption and emissions produced; (ii) the trade-off between accuracy and energy efficiency; and (iii) the difference on the method of measurement of the energy consumed using software-based and hardware-based tools.

show abstract

An Approach for Weed Detection Using CNNs And Transfer Learning

Cited by 8 publications

References 18 publications

A Deep Learning Model Compression and Ensemble Approach for Weed Detection

A Deep Learning Model Compression and Ensemble Approach for Weed Detection

Weed Detection Using Deep Learning: A Systematic Literature Review

Energy Efficiency of Training Neural Network Architectures: An Empirical Study

Contact Info

Product

Resources

About