Study of Deep Learning-based models for Single Image Super-Resolution

Soufi, Omar; Belouadha, Fatima-Zahra

doi:10.18280/ria.360616

Cited by 4 publications

(2 citation statements)

References 74 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As we can see in Table 5, our proposed method outperforms all other method and has training time less than 95 second unlike all compared methods. [33] 159.14 7 30 Resnet-152 [34] 1281 (21 minutes20 seconds) 7 30 VGG-19 [35] 8.77 7 30 AlexNet [36] 850 (14 minutes 10 seconds) 7 30 GoogLeNet (Inception V1) [37] 3716.42 (61.94 minutes) 7 30 Inception V 3 [38] 825.87 (14 minutes) 7 30 VGG 16 [39] 913.73 (15 minutes) 7 30 DenseNet121 [40] 791.99 (13 minutes) 7 30 SqueezeNet [41] 152.23 7 30…”

Section: Comparison Of Our Methods With Other Architecturementioning

confidence: 99%

Deep Learning and Machine Learning Based Method for Crop Disease Detection and Identification Using Autoencoder and Neural Network

Boukhris,

Jilali,

Asri

2024

RIA

View full text Add to dashboard Cite

Crop diseases present a major threat to agricultural output, disrupting both the quantity and quality of production. Disease diagnosis remains a challenge for farmers, primarily due to limited knowledge and the need for specialized agricultural engineering expertise. To solve these problems, a new technique named the Autoencoder Latent Space-Neural Network (ALS-NN) was introduced in this study. It combines the strengths of autoencoders and neural networks to find crop diseases. Data processing is the first step of the methodology, and then data compression into a latent space follows. This compressed data serves as the input for the neural network, facilitating efficient crop disease classification. This approach capitalizes on the autoencoder's capacity for dimensionality reduction, data compression, and encoding, which is particularly beneficial when handling high-dimensional data. The reduced data dimensionality enables the neural network to process the information more efficiently. The ALS-NN model, by compressing data, focuses on the crucial information for the classification process, thereby enhancing computational speed and reducing the number of trained parameters. This results in time efficiencies during disease detection operations, mitigating the detrimental effects of diseases on crop yields. The integration of autoencoders and neural networks forms a potent strategy for disease detection, leveraging the autoencoders' capabilities for dimensionality reduction, anomaly detection, and feature learning, coupled with the classification and generalization abilities of neural networks. This hybrid approach can potentially lead to more precise, efficient, and interpretable disease detection system. PlantVillage is used with 10 crop types. We used the first part of autoencoder (The encoder) to compress images into Latent space; for classification, the result is subsequently fed into a neural network. Our model (ALS-NN) achieved 90% for test accuracy and 90% for validation accuracy.

show abstract

Section: Comparison Of Our Methods With Other Architecturementioning

confidence: 99%

Deep Learning and Machine Learning Based Method for Crop Disease Detection and Identification Using Autoencoder and Neural Network

Boukhris,

Jilali,

Asri

2024

RIA

View full text Add to dashboard Cite

show abstract

“…The emergence of deep learning in recent decade has revolutionized the study of image processing and computer vision [11,12]. In this field, automatic depth estimation is now a research hot spot that has brought unprecedented opportunities for animation production.…”

Section: Introductionmentioning

confidence: 99%

Automatic Depth Estimation and Background Blurring of Animated Scenes Based on Deep Learning

He,

Jia

2023

View full text Add to dashboard Cite

Animation technology enables more accurate depth estimation and background blurring of animated scenes as it can enhance the sense of reality of the vision and increase its depth, thus it has become a hot spot in relevant research and production these days. However, although deep learning has made significant progresses in many research fields, its application in depth estimation and background blurring of animated scenes is still facing a few challenges. Most available technologies are for real world images, not animations, so there are certain difficulties capturing the unique styles of animations and their details. This study proposes two technical schemes specifically designed for animated scenes: a depth estimation model based on DenseNet, and a deblurring algorithm based on Very Deep Super Resolution (VDSR), in the hopes of providing solutions for the above mentioned matters, as well as forging more efficient and accurate tools for the animation industry.

show abstract

Image super-resolution based on improved ESRGAN and its application in camera calibration

Chen,

Li,

Yao

et al. 2025

Measurement

View full text Add to dashboard Cite

Study of Deep Learning-based models for Single Image Super-Resolution

Cited by 4 publications

References 74 publications

Deep Learning and Machine Learning Based Method for Crop Disease Detection and Identification Using Autoencoder and Neural Network

Deep Learning and Machine Learning Based Method for Crop Disease Detection and Identification Using Autoencoder and Neural Network

Automatic Depth Estimation and Background Blurring of Animated Scenes Based on Deep Learning

Image super-resolution based on improved ESRGAN and its application in camera calibration

Contact Info

Product

Resources

About