Abstract
Image recognition based on convolutional neural networks (CNNs) has recently been shown to deliver the state-of-the-art performance in various areas of computer vision and image processing. Nevertheless, applying a deep CNN to no-reference image quality assessment (NR-IQA) remains a challenging task due to critical obstacles, i.e., the lack of a training database. In this paper, we propose a CNN-based NR-IQA framework that can effectively solve this problem. The proposed method - deep image quality assessor (DIQA) - separates the training of NR-IQA into two stages: 1) an objective distortion part and 2) a human visual system-related part. In the first stage, the CNN learns to predict the objective error map, and then the model learns to predict subjective score in the second stage. To complement the inaccuracy of the objective error map prediction on the homogeneous region, we also propose a reliability map. Two simple handcrafted features were additionally employed to further enhance the accuracy. In addition, we propose a way to visualize perceptual error maps to analyze what was learned by the deep CNN model. In the experiments, the DIQA yielded the state-of-the-art accuracy on the various databases.
| Original language | English |
|---|---|
| Article number | 8383698 |
| Pages (from-to) | 11-24 |
| Number of pages | 14 |
| Journal | IEEE Transactions on Neural Networks and Learning Systems |
| Volume | 30 |
| Issue number | 1 |
| DOIs | |
| Publication status | Published - 2019 Jan |
Fingerprint
All Science Journal Classification (ASJC) codes
- Software
- Computer Science Applications
- Computer Networks and Communications
- Artificial Intelligence
Cite this
}
Deep CNN-Based Blind Image Quality Predictor. / Kim, Jongyoo; Nguyen, Anh Duc; Lee, Sanghoon.
In: IEEE Transactions on Neural Networks and Learning Systems, Vol. 30, No. 1, 8383698, 01.2019, p. 11-24.Research output: Contribution to journal › Article
TY - JOUR
T1 - Deep CNN-Based Blind Image Quality Predictor
AU - Kim, Jongyoo
AU - Nguyen, Anh Duc
AU - Lee, Sanghoon
PY - 2019/1
Y1 - 2019/1
N2 - Image recognition based on convolutional neural networks (CNNs) has recently been shown to deliver the state-of-the-art performance in various areas of computer vision and image processing. Nevertheless, applying a deep CNN to no-reference image quality assessment (NR-IQA) remains a challenging task due to critical obstacles, i.e., the lack of a training database. In this paper, we propose a CNN-based NR-IQA framework that can effectively solve this problem. The proposed method - deep image quality assessor (DIQA) - separates the training of NR-IQA into two stages: 1) an objective distortion part and 2) a human visual system-related part. In the first stage, the CNN learns to predict the objective error map, and then the model learns to predict subjective score in the second stage. To complement the inaccuracy of the objective error map prediction on the homogeneous region, we also propose a reliability map. Two simple handcrafted features were additionally employed to further enhance the accuracy. In addition, we propose a way to visualize perceptual error maps to analyze what was learned by the deep CNN model. In the experiments, the DIQA yielded the state-of-the-art accuracy on the various databases.
AB - Image recognition based on convolutional neural networks (CNNs) has recently been shown to deliver the state-of-the-art performance in various areas of computer vision and image processing. Nevertheless, applying a deep CNN to no-reference image quality assessment (NR-IQA) remains a challenging task due to critical obstacles, i.e., the lack of a training database. In this paper, we propose a CNN-based NR-IQA framework that can effectively solve this problem. The proposed method - deep image quality assessor (DIQA) - separates the training of NR-IQA into two stages: 1) an objective distortion part and 2) a human visual system-related part. In the first stage, the CNN learns to predict the objective error map, and then the model learns to predict subjective score in the second stage. To complement the inaccuracy of the objective error map prediction on the homogeneous region, we also propose a reliability map. Two simple handcrafted features were additionally employed to further enhance the accuracy. In addition, we propose a way to visualize perceptual error maps to analyze what was learned by the deep CNN model. In the experiments, the DIQA yielded the state-of-the-art accuracy on the various databases.
UR - http://www.scopus.com/inward/record.url?scp=85048558863&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85048558863&partnerID=8YFLogxK
U2 - 10.1109/TNNLS.2018.2829819
DO - 10.1109/TNNLS.2018.2829819
M3 - Article
C2 - 29994270
AN - SCOPUS:85048558863
VL - 30
SP - 11
EP - 24
JO - IEEE Transactions on Neural Networks and Learning Systems
JF - IEEE Transactions on Neural Networks and Learning Systems
SN - 2162-237X
IS - 1
M1 - 8383698
ER -