The Optical Subsystem for the Empty Containers Recognition and Sorting in a Reverse Vending Machine

Kokoulin, Andrey N.; Kiryanov, D.A.

doi:10.23919/splitech.2019.8782990

Cited by 12 publications

(6 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Kokoulin et al [12], Kokoulin and Kiryanov [13] introduced an image-based object classification system for PET and aluminum cans. They compared various CNN models such as LeNet, AlexNet, and SqueezeNet, and showed that the best results are derived when LeNet classifies an object into two classes, PET bottles and cans.…”

Section: Rvm and Cnnmentioning

confidence: 99%

“…As illustrated in Figure 3, our dataset consists of the top view and the front view images subsets. We further divided each subset into PET, Can, and Non-target classes, which are the most common classification classes of the commercial RVMs [39,40] and the previous studies in [12,13]. The objects belonging to each class are summarized in Table 1.…”

Section: Datasetmentioning

confidence: 99%

“…We propose an ensemble CNN model using the top and front view images to resolve the limitations of previous state-of-the-art RVM studies using a single image-based system [12,13] or multi-modal network [14]. The main goal of our ensemble approach is to classify fraud objects with high accuracy on a vision-only system.…”

Section: The Proposed Cnn Ensemble Modelmentioning

confidence: 99%

“…Therefore, many studies have been conducted to apply CNNs for the RVM system. For example, Kokoulin et al [12], Kokoulin and Kiryanov [13] compared the performance of single image-based systems with multiple CNNs to classify input objects. Park et al [14] showed that a single image-based system cannot accurately identify intended fraud objects and proposed a multi-modal network that uses various modalities such as image, weight, and ultrasound.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Dual Image-Based CNN Ensemble Model for Waste Classification in Reverse Vending Machine

2021

View full text Add to dashboard Cite

A reverse vending machine motivates citizens to bring recyclable waste by rewarding them, which is a viable solution to increase the recycling rate. Reverse vending machines generally use near-infrared sensors, barcode sensors, or cameras to classify recycling resources. However, sensor-based reverse vending machines suffer from a high configuration cost and the limited scope of target objects, and conventional single image-based reverse vending machines usually make erroneous predictions about intentional fraud objects. This paper proposes a dual image-based convolutional neural network ensemble model to address these problems. For this purpose, we first created a prototype reverse vending machine and constructed an image dataset containing two cross-sections of objects, top and front view. Then, we chose convolutional neural network models widely used in image classification as the candidates for building an accurate and lightweight ensemble model. Considering the size and classification performance of candidates, we constructed the best-fit ensemble combination and evaluated its classification performance. The final ensemble model showed a classification accuracy higher than 95% for all target classes, including fraud objects. This result proves that our approach achieves better robustness against intentional fraud objects than single image-based models and thus can broaden the scope for target resources. The measurement results on lightweight embedded platforms also demonstrated that our model provides a short inference time that is enough to facilitate the real-time execution of reverse vending machines based on low-cost edge artificial intelligence devices.

show abstract

Section: Rvm and Cnnmentioning

confidence: 99%

Section: Datasetmentioning

confidence: 99%

Section: The Proposed Cnn Ensemble Modelmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Dual Image-Based CNN Ensemble Model for Waste Classification in Reverse Vending Machine

2021

View full text Add to dashboard Cite

show abstract

“…In average the CNN model can produce more than 85% accuracy. They later on test the module in real implementation by combining weigh sensor with the CNN for fraud detection [7].…”

Section: Introductionmentioning

confidence: 99%

Reverse Vending Machine Item Verification Module using Classification and Detection Model of CNN

Tomari¹,

Razali²,

Santosa³

et al. 2021

IJACSA

View full text Add to dashboard Cite

Photorealism versus photography. AI-generated depiction in the age of visual disinformation

Hausken

2024

Journal of Aesthetics & Culture

View full text Add to dashboard Cite

The Optical Subsystem for the Empty Containers Recognition and Sorting in a Reverse Vending Machine

Cited by 12 publications

References 5 publications

Dual Image-Based CNN Ensemble Model for Waste Classification in Reverse Vending Machine

Dual Image-Based CNN Ensemble Model for Waste Classification in Reverse Vending Machine

Reverse Vending Machine Item Verification Module using Classification and Detection Model of CNN

Photorealism versus photography. AI-generated depiction in the age of visual disinformation

Contact Info

Product

Resources

About