Multi-source domain adaptation for quality control in retail food packaging

Thota, Mamatha; Kollias, Stefanos; Swainson, Mark; Leontidis, Georgios

doi:10.1016/j.compind.2020.103293

Cited by 22 publications

(16 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The use of machine learning within agri-food has seen successful application in yield prediction [17], crop disease detection [18], demand prediction [19], and production safety [20]. We suggest that many of these already impactful advancements can be replicated (or at least approximated) without direct data pooling, thus making them much more widely usable.…”

Section: Related Workmentioning

confidence: 99%

The Role of Cross-Silo Federated Learning in Facilitating Data Sharing in the Agri-Food Sector

Durrant¹,

Marković²,

May³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Data sharing remains a major hindering factor when it comes to adopting emerging AI technologies in general, but particularly in the agri-food sector. Protectiveness of data is natural in this setting: data is a precious commodity for data owners, which if used properly can provide them with useful insights on operations and processes leading to a competitive advantage. Unfortunately, novel AI technologies often require large amounts of training data in order to perform well, something that in many scenarios is unrealistic. However, recent machine learning advances, e.g. federated learning and privacy-preserving technologies, can offer a solution to this issue via providing the infrastructure and underpinning technologies needed to use data from various sources to train models without ever sharing the raw data themselves. In this paper, we propose a technical solution based on federated learning that uses decentralized data, (i.e. data that are not exchanged or shared but remain with the owners) to develop a cross-silo machine learning model that facilitates data sharing across supply chains. We focus our data sharing proposition on improving production optimization through soybean yield prediction, and provide potential use-cases that such methods can assist in other problem settings. Our results demonstrate that our approach not only performs better than each of the models trained on an individual data source, but also that data sharing in the agri-food sector can be enabled via alternatives to data exchange, whilst also helping to adopt emerging machine learning technologies to boost productivity.

show abstract

Section: Related Workmentioning

confidence: 99%

The Role of Cross-Silo Federated Learning in Facilitating Data Sharing in the Agri-Food Sector

Durrant¹,

Marković²,

May³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Over the last few years, Deep Learning (DL) [22] has been successfully applied across numerous applications and domains due to the availability of large amounts of labeled data, such as computer vision and image processing [34,42,37,8], signal processing [2,33,15], autonomous driving [26,41,11], agri-food technologies [1,20], medical imaging [19,25], etc. Most of the applications of DL techniques, such as the aforementioned ones, refer to supervised learning, it requires manually labeling a dataset, which is a very time consuming, cumbersome and expensive process that has led to the widespread use of certain datasets, e.g.…”

Section: Introductionmentioning

confidence: 99%

Contrastive Domain Adaptation

Thota

Leontidis

2021

2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)

Self Cite

View full text Add to dashboard Cite

Recently, contrastive self-supervised learning has become a key component for learning visual representations across many computer vision tasks and benchmarks. However, contrastive learning in the context of domain adaptation remains largely underexplored. In this paper, we propose to extend contrastive learning to a new domain adaptation setting, a particular situation occurring where the similarity is learned and deployed on samples following different probability distributions without access to labels. Contrastive learning learns by comparing and contrasting positive and negative pairs of samples in an unsupervised setting without access to source and target labels. We have developed a variation of a recently proposed contrastive learning framework that helps tackle the domain adaptation problem, further identifying and removing possible negatives similar to the anchor to mitigate the effects of false negatives. Extensive experiments demonstrate that the proposed method adapts well, and improves the performance on the downstream domain adaptation task.

show abstract

“…Medus et al (2021) 11 , for instance, proposed a system which, by using a convolutional neural network (CNN) as a classifier in heat-sealed food trays, is able to automatically detect anomalies during the packaging process in order to discard the faulty tray and avoid human consumption. Thota et al (2020) 12 presented a multi-source deep-learning-based domain adaptation system to identify and verify the presence and legibility of use-by date information from food packaging photos taken as part of the validation process as the products pass along the food production line. Brunelli et al (2019) 13 designed a deep-learning-based approach for production performance forecasting in fresh products packaging.…”

Section: Introductionmentioning

confidence: 99%

“…SVM = Support Vector Machine, RBF = Radial Basis Function, CNN = Convolutional Neural Network, Pre = Pre-trained with ImageNet, HSI = Hyperspectral Imaging, PL = Polarized, LS = Laser Scatter, ch = channels, f = features, Mono = Monochrome, and IRAS = Infrared image Acquisition System. Medus et al 11 Thota et al 12 Prada-López et al 17 Xie et al 18 Barnes et al 40 Our method Quality control Sealing of heat-sealed food trays (Anomalies detected: washers, sugar, flange, plywood, cork, elastic rubber, wood, paper, sliver paper, hair, and polarized plastic) Presence and legibility of use-by date information from food packages Types of coffee and adulterations Atlantic salmon bone residues Detect copper wire in the seal of heat-sealed packages Sealing of thermoforming food packages Method (Accuracy) SVM RBF 84.3% CNN-ResNet18 96% CNN-ResNet18(Pre) 96% CNN-ResNet50 95% 94% CNN-ResNet50(Pre) 92.5% 97% CNN-ResNet34(Pre) 98.6% CNN-VGG16...…”

Section: Introductionmentioning

confidence: 99%

Deep learning for the quality control of thermoforming food packages

Banús

Boada

Xiberta

et al. 2021

Sci Rep

View full text Add to dashboard Cite

Quality control is a key process designed to ensure that only products satisfying the defined quality requirements reach the end consumer or the next step in a production line. In the food industry, in the packaging step, there are many products that are still evaluated by human operators. To automate the process and improve efficiency and effectiveness, computer vision and artificial intelligence techniques can be applied. This automation is challenging since specific strategies designed according to the application scenario are required. Focusing on the quality control of the sealing and closure of matrix-shaped thermoforming food packages, the aim of the article is to propose a deep-learning-based solution designed to automatically perform the quality control while satisfying production cadence and ensuring 100% inline inspection of the products. Particularly, the designed computer vision system and the image-based criteria defined to determine when a product has to be accepted or rejected are presented. In addition, the vision control software is described with special emphasis on the different convolutional neural network (CNN) architectures that have been considered (ResNet18, ResNet50, Vgg19 and DenseNet161, non-pre-trained and pre-trained on ImageNet) and on the specifically designed dataset. To test the solution, different experiments are carried out in the laboratory and also in a real scenario, concluding that the proposed CNN-based approach improves the efficiency and security of the quality control process. Optimal results are obtained with the pre-trained DenseNet161, achieving false positive rates that range from 0.03 to 0.30% and false negative rates that range from 0 to 0.07%, with a rejection rate between 0.64 and 5.09% of production, and being able to detect at least 99.93% of the sealing defects that occur in any production. The modular design of our solution as well as the provided description allow it to adapt to similar scenarios and to new deep-learning models to prevent the arrival of faulty products to end consumers by removing them from the automated production line.

show abstract

Multi-source domain adaptation for quality control in retail food packaging

Cited by 22 publications

References 28 publications

The Role of Cross-Silo Federated Learning in Facilitating Data Sharing in the Agri-Food Sector

The Role of Cross-Silo Federated Learning in Facilitating Data Sharing in the Agri-Food Sector

Contrastive Domain Adaptation

Deep learning for the quality control of thermoforming food packages

Contact Info

Product

Resources

About