Evaluating Convolutional Neural Networks for No -Reference Image Quality Assessment

Apostolidis, Kyriakos D.; Polyzos, Theodore; Grigoriadis, Ioannis; Papakostas, George A.

doi:10.1109/icspis53734.2021.9652176

Cited by 3 publications

(1 citation statement)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The evolution of deep learning and computer hardware has helped computer vision applications become reality. Some disciplines that use DL for computer vision tasks are robotics [ 1 ], image quality assessment [ 2 ], biometrics [ 3 ], face recognition [ 4 ], image classification [ 5 ], autonomous vehicles [ 6 ], etc. One of the most important applications in CV is medical image analysis, where usually DL models were trained to diagnose or predict several diseases from numerous modalities such as MRI, CT-scans, X-rays, Histopathology images, etc.…”

Section: Introductionmentioning

confidence: 99%

Digital Watermarking as an Adversarial Attack on Medical Image Analysis with Deep Learning

Apostolidis

Papakostas

2022

J. Imaging

Self Cite

View full text Add to dashboard Cite

In the past years, Deep Neural Networks (DNNs) have become popular in many disciplines such as Computer Vision (CV), and the evolution of hardware has helped researchers to develop many powerful Deep Learning (DL) models to deal with several problems. One of the most important challenges in the CV area is Medical Image Analysis. However, adversarial attacks have proven to be an important threat to vision systems by significantly reducing the performance of the models. This paper brings to light a different side of digital watermarking, as a potential black-box adversarial attack. In this context, apart from proposing a new category of adversarial attacks named watermarking attacks, we highlighted a significant problem, as the massive use of watermarks, for security reasons, seems to pose significant risks to vision systems. For this purpose, a moment-based local image watermarking method is implemented on three modalities, Magnetic Resonance Images (MRI), Computed Tomography (CT-scans), and X-ray images. The introduced methodology was tested on three state-of-the art CV models, DenseNet 201, DenseNet169, and MobileNetV2. The results revealed that the proposed attack achieved over 50% degradation of the model’s performance in terms of accuracy. Additionally, MobileNetV2 was the most vulnerable model and the modality with the biggest reduction was CT-scans.

show abstract

Section: Introductionmentioning

confidence: 99%

Digital Watermarking as an Adversarial Attack on Medical Image Analysis with Deep Learning

Apostolidis

Papakostas

2022

J. Imaging

Self Cite

View full text Add to dashboard Cite

show abstract

Visual Quality Inspection of Pomegranate Crop Using a Novel Dataset and Deep Learning

Koufatzis

Vrochidou

Papakostas

2022

2022 29th International Conference on Systems, Signals and Image Processing (IWSSIP)

View full text Add to dashboard Cite

Research on objective assessment of dual-band color night vision fusion image quality based on scene understanding

Gao,

Yi,

Tian

et al. 2023

Opt. Eng.

View full text Add to dashboard Cite

The color quality and sharpness of the dual-band color fusion images have the greatest impact on the audience's understanding of the scene content. Traditional image quality assessment (IQA) methods focus on the luminance contrast of image and attach limited attention to pixel-wise relationship. As a result, these methods perform poorly on color-related distortions. A objective assessment model for the comprehensive quality of fusion images is proposed. According to the correlation between the gray level co-occurrence matrix (GLCM) and the image color, the model extracts the homogeneity, angular second moment, and contrast of the GLCM as the image color quality features. Gradient-weighted class activation mapping is used to calculate the heat map, reflecting the visual saliency of each region. We use the visual saliency weight of each region to weight the rotation and uniform invariant local binary pattern histograms of the corresponding regions to obtain sharpness features. The experimental results show that the performance of the proposed method is better than the existing eight no-reference IQA models.

show abstract

Evaluating Convolutional Neural Networks for No -Reference Image Quality Assessment

Cited by 3 publications

References 18 publications

Digital Watermarking as an Adversarial Attack on Medical Image Analysis with Deep Learning

Digital Watermarking as an Adversarial Attack on Medical Image Analysis with Deep Learning

Visual Quality Inspection of Pomegranate Crop Using a Novel Dataset and Deep Learning

Research on objective assessment of dual-band color night vision fusion image quality based on scene understanding

Contact Info

Product

Resources

About