Open Access Open Access  Restricted Access Subscription Access

A Review: Multi-Expert Convolutional Neural Networks for Not Included Image Quality Assessment

Kanisha Patel

Abstract


It is possible to assess the quality of distorted images without having to consult the original image by using the No Reference (NR) Picture Quality Evaluation (IQA) calculation. This attribute is important in the field of handling pictures. There is difficulty in the current NR IQA calculations to keep up with the best exhibition due to the variety of contortion types and picture contents. To address this problem, we develop an innovative NR IQA calculation considering multi-master convolutional brain organizations (CNNs), which includes combination calculation, bending type grouping, and IQA calculations considering CNNs. A distortion type classifier first detects the different forms of distortion present in the input image. Next, provide an IQA computation based on multi-master CNN for each of these mutilation types. Finally, a combined computation is performed to sum together the effects of both the multi-master CNN-based picture quality forecasts and the contortion types.


Full Text:

PDF

References


Ma, Y., Cai, X., & Sun, F. (2020). Towards No-Reference Image Quality Assessment Based on Multi-Scale Convolutional Neural Network. CMES-Computer Modeling in Engineering & Sciences, 123(1).

de Maissin, A., Vallée, R., Flamant, M., Fondain-Bossiere, M., Le Berre, C., Coutrot, A., ... & Bourreille, A. (2021). Multi-expert annotation of Crohn’s disease images of the small bowel for automatic detection using a convolutional recurrent attention neural network. Endoscopy International Open, 9(07), E1136-E1144.

Yang, J., Liu, T., Jiang, B., Song, H., & Lu, W. (2018). 3D panoramic virtual reality video quality assessment based on 3D convolutional neural networks. IEEE Access, 6, 38669-38682.

Ravela, R., Shirvaikar, M., & Grecos, C. (2019, May). No-reference image quality assessment based on deep convolutional neural networks. In Real-Time Image Processing and Deep Learning 2019 (Vol. 10996, p. 1099604). SPIE.


Refbacks

  • There are currently no refbacks.