No Reference Image Quality Assessment for JPEG Images using Machine Learning Approach
Jayashri V. Bagade, Yogesh Dandawate, Kulbir Singh · 2021
Popular and widely used JPEG images suffer from blocking artifact. Blocking artifact degrades the quality of an image. No Reference Image Quality Assessment (NRIQA) metric using block based features are designed in this paper. Artificial Neural Network (ANN) and Neuro Fuzzy Classifier (NFC) are employed to evaluate the performance of the proposed metric. It is tested for LIVE, TID2008 and TID2013 datasets. Results show improvement in accuracy independent of image databases. The proposed metric is compared with state-of-the-art metrics for LIVE dataset.