A Non Reference Based Image Quality Predictor Using Deep Convolutional Neural Network: A Review

Pravitha K.

Abstract


Now days the multimedia technology is on the way of rapid development .So any technology which can predict the image quality has importance. There are different technologies available for the Image Quality Assessment (IQA). Each technology differs from one another. The Convolution Neural Network (CNN)is using in computer vision and  image processing  .So for Image Quality  Assessment  we can use Convolution Neural Network. For Image Quality Assessment there is a subjective and objective method. The objective methods divided in to, Full-Reference IQA (FR-IQA), Reduced-Reference IQA (RR-IQA), Non-Reference IQA (NR-IQA). In all situations it is not possible to have a reference image so go for Non-Reference IQA (NR-IQA). In this proposed method – the steps are first performing normalization as preprocessing on image and then the training process. The training process of NR-IQA here divided in to 1) Objective distortion part and 2) Human visual system related part. In this the First step is predicting the objective error map, then model learn to predict the subjective score. The reliability map prediction used as the inaccuracy solver of the objective error map prediction on the homogeneous region, some handcrafted features used to further enhance the accuracy. Also comparing the performance using different training methods.

 

Keywords: IQA, CNN, FR-IQA, NR-IQA, RR-IQA

 


References


Moorthy A.K., Bovik A.C. Blind image quality assessment: From natural scene statistics to perceptual quality. IEEE Trans. Image Process. vol. 20, no. 12, 3350–3364p, Dec. 2011.

Kang L., Ye P., Li Y., Doermann D. Convolutional neural networks for no-reference image quality assessment. In Proc. IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR). Jun. 2014, 1733–1740p.

Ma K., Liu W., Liu T., Wang Z., Tao D. dipIQ: Blind image quality assessment by learning-to-rank discriminable image pairs. IEEE Trans. Image Process. vol. 26, no. 8, 3951–3964p, Aug. 2017.

Ma K., Liu W., Zhang K., et al. End-to-End blind image quality assessment using deep neural networks. IEEE Trans. Image Process. vol. 27, no. 3, 1202–1213p, Mar. 2018.


Refbacks

  • There are currently no refbacks.