Active learning with MaskAL reduces annotation effort for training Mask R-CNN on a broccoli dataset with visually similar classes

Blok, Pieter M.; Kootstra, Gert; Elghor, Hakim Elchaoui; Diallo, Boubacar; Evert, F.K. van; Henten, E.J. van

doi:10.1016/j.compag.2022.106917

Cited by 21 publications

(11 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Object detection methods can be divided into two categories: two stage and one stage. In the two-stage object detection, the objects are first localized and then classified, and the representative algorithms are R-CNN [11], Fast R-CNN, and R-FCN [12]. One-stage object detection regards object detection as a regression problem and performs localization and classification at the same time.…”

Section: Related Workmentioning

confidence: 99%

“…After determining the depth of the Backbone network, on the basis of YOLOv5_16, Backbone adds the SPP pyramid pooling module, and Neck uses the FPN and PAN structure to fuse the features of the 8-fold and 16-fold downsampling layers, naming the model YOLOv5s_B. In order to explore the optimal pooling effect of SPP, four common SPP pooling kernels are tested in this paper: (3, 5, 7), (5,7,9), (7,9,13) and (9,11,13), named YOLOv5s_B_a, YOLOv5s_B_b, YOLOv5s_B_c, and YOLOv5s_B_d, respectively. Its structure is shown in Figure 9.…”

Section: Influence Of Feature Fusion On Detection Performancementioning

confidence: 99%

See 1 more Smart Citation

Research on Fingerprint Security Based on Improved Yolo Algorithm

2022

Mobile Information Systems

View full text Add to dashboard Cite

The quantitative identification technology based on the statistical law of fingerprint features has become a new research difficulty and focus, and the automatic detection and classification of fingerprint features are the basis for realizing automatic fingerprint feature statistics. In this paper, a YOLO-based fingerprint feature detection method was proposed. First, a fingerprint feature dataset was established, which contained a total of 4,000 annotated fingerprint images; then, according to the characteristics of small size and dense distribution of fingerprint feature points, the YOLO network structure was improved, the original large target feature detection layer by 32-fold downsampling was deleted, and a new small feature fusion layer was added; the FPN, PAN, and SPP structures were used to achieve local and global feature extraction through multiple-scale fusion methods; finally, the SE channel attention mechanism module was added to effectively enhance the model robustness and detection ability of dense small objects. The experimental results show that compared with the improved FP-YOLO model of the original model, when the detection speed is basically unchanged, the mAP0.5 value is increased from 93.0% to 97.4%, and the weight is reduced by 3/4.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Influence Of Feature Fusion On Detection Performancementioning

confidence: 99%

Research on Fingerprint Security Based on Improved Yolo Algorithm

2022

Mobile Information Systems

View full text Add to dashboard Cite

show abstract

“…The pretreatment focused on conversions of full image datasets from RGB to four color spaces, preserving details and maintaining a balanced number of elements per group, which helps in the proper training of the neural network ( Ciocca, Napoletano & Schettini, 2018 ; Blok et al, 2022 ). This is similar to what was performed by Castro et al (2019a) who, using RGB images, extracted mean values and converted them to another L*a*b* and HSV in a previous stage of classifiers training, obtaining that the best combination was supported vector machine and RGB color space.…”

Section: Resultsmentioning

confidence: 99%

“…In the agricultural field, CNNs have been successfully applied in crop variety identification ( Too et al, 2019 ), haploid and diploid seeds ( Altuntaş, Cömert & Kocamaz, 2019 ), nematodes ( Abade et al, 2022 ), plant disease recognition ( Too et al, 2019 ), damage in milled rice grains ( Moses et al, 2022 ), broccoli head quality discrimination ( Blok et al, 2022 ), crop pest ( Ayan, Erbay & Varçın, 2020 ), microstructural elements discrimination ( Castro et al, 2019b ), characterization of emulsions ( Lu et al, 2021 ), among others. Using different common CNNs architectures (AlexNet, resNet, MobileNet, Inception, VGG16, DenseNet, among others), new architectures, and-or training approaches.…”

Section: Introductionmentioning

confidence: 99%

Discrimination of foreign bodies in quinoa (Chenopodium quinoa Willd.) grains using convolutional neural networks with a transfer learning approach

Ávila-George

De-la-Torre

Sánchez-Garcés

et al. 2023

PeerJ

View full text Add to dashboard Cite

The rising interest in quinoa (Chenopodium quinoa Willd.) is due to its high protein content and gluten-free condition; nonetheless, the presence of foreign bodies in quinoa processing facilities is an issue that must be addressed. As a result, convolutional neural networks have been adopted, mostly because of their data extraction capabilities, which had not been utilized before for this purpose. Consequently, the main objective of this work is to evaluate convolutional neural networks with a learning transfer for foreign bodies identification in quinoa samples. For experimentation, quinoa samples were collected and manually split into 17 classes: quinoa grains and 16 foreign bodies. Then, one thousand images were obtained from each class in RGB space and transformed into four different color spaces (L*a*b*, HSV, YCbCr, and Gray). Three convolutional neural networks (AlexNet, MobileNetv2, and DenseNet-201) were trained using the five color spaces, and the evaluation results were expressed in terms of accuracy and F-score. All the CNN approaches compared showed an F-score ranging from 98% to 99%; both color space and CNN structure were found to have significant effects on the F-score. Also, DenseNet-201 was the most robust architecture and, at the same time, the most time-consuming. These results evidence the capacity of CNN architectures to be used for the discrimination of foreign bodies in quinoa processing facilities.

show abstract

“…Most current agro-food robotics research aims to improve the perception capabilities of the systems, which increases the robots' understanding of a scene. In recent years, most of the research has been done in increasing detection and localization performance, especially through deep learning (Blok et al, 2022;Ruigrok et al, 2020). Although these methods have shown to be able to deal with variation to some extent, the challenge of occlusion is not adequately addressed (Montoya-Cavero et al, 2022;Kootstra et al, 2021).…”

Section: Introductionmentioning

confidence: 99%

Development and evaluation of automated localization and reconstruction of all fruits on tomato plants in a greenhouse based on multi-view perception and 3D multi-object tracking

Rincon¹,

Henten²,

Kootstra³

2022

Preprint

View full text Add to dashboard Cite

Accurate representation and localization of relevant objects is important for robots to perform tasks. Building a generic representation that can be used across different environments and tasks is not easy, as the relevant objects vary depending on the environment and the task. Furthermore, another challenge arises in agro-food environments due to their complexity, and high levels of clutter and occlusions. In this paper, we present a method to build generic representations in highly occluded agro-food environments using multi-view perception and 3D multi-object tracking. Our representation is built upon a detection algorithm that generates a partial point cloud for each detected object. The detected objects are then passed to a 3D multi-object tracking algorithm that creates and updates the representation over time. The whole process is performed at a rate of 10 Hz. We evaluated the accuracy of the representation on a real-world agro-food environment, where it was able to successfully represent and locate tomatoes in tomato plants despite a high level of occlusion. We were able to estimate the total count of tomatoes with a maximum error of 5.08% and to track tomatoes with a tracking accuracy up to 71.47%. Additionally, we showed that an evaluation using tracking metrics gives more insight in the errors in localizing and representing the fruits.

show abstract

Active learning with MaskAL reduces annotation effort for training Mask R-CNN on a broccoli dataset with visually similar classes

Cited by 21 publications

References 15 publications

Research on Fingerprint Security Based on Improved Yolo Algorithm

Research on Fingerprint Security Based on Improved Yolo Algorithm

Discrimination of foreign bodies in quinoa (Chenopodium quinoa Willd.) grains using convolutional neural networks with a transfer learning approach

Development and evaluation of automated localization and reconstruction of all fruits on tomato plants in a greenhouse based on multi-view perception and 3D multi-object tracking

Contact Info

Product

Resources

About