International Conference on Computer and Knowledge Engineering

Home / 12th International Conference on Computer and Knowledge Engineering

No-Reference Video Quality Assessment by Deep Feature Maps Relations

Authors :

Amir Hossein Bakhtiari¹ Azadeh Mansouri²

1- Department of Electrical and Computer Engineering Faculty of Engineering Kharazmi University Tehran, Iran 2- Department of Electrical and Computer Engineering Faculty of Engineering Kharazmi University Tehran, Iran

Keywords :

gram matrix،no-reference video quality assessment،convolutional neural network،human visual system

Abstract :

The aim of blind video quality assessment (BVQA) methods is to evaluate the perceptual quality of a distorted video without any prior information regarding its reference one. Numerous deep network-based techniques have so far been introduced. These methods often pool the features obtained for each frame in different ways to generate a video representation and evaluate the quality. In this paper, we introduce a novel technique for obtaining frame-level features to assess quality. In order to accomplish this, we explored the deep feature maps relations as useful information for video quality assessment. The Gram Matrix generated in each layer is analyzed and explored as higher-order quality features using pre-trained networks. The deep feature relations can be considered similar to the covariance matrix, which indicates the correlations between different feature maps. In fact, these correlations reflect the structural information of each frame. After features extraction and pooling, support vector regression (SVR) is adopted in order to provide a quality score. Experimental results show the effectiveness of the proposed features and offer comparable performance to the state-of-the-art methods.