Abstract
This paper proposes a method of accurately assessing image quality without a reference image by using a deep convolutional neural network. Existing training based methods usually utilize a compact set of linear filters for learning features of images captured by different sensors to assess their quality. These methods may not be able to learn the semantic features that are intimately related with the features used in human subject assessment. Observing this drawback, this work proposes training a deep convolutional neural network (CNN) with labelled images for image quality assessment. The ReLU in the CNN allows non-linear transformations for extracting high-level image features, providing a more reliable assessment of image quality than linear filters. To enable the neural network to take images of any arbitrary size as input, the spatial pyramid pooling (SPP) is introduced connecting the top convolutional layer and the fully-connected layer. In addition, the SPP makes the CNN robust to object deformations to a certain extent. The proposed method taking an image as input carries out an end-to-end learning process, and outputs the quality of the image. It is tested on public datasets. Experimental results show that it outperforms existing methods by a large margin and can accurately assess the image quality on images taken by different sensors of varying sizes.
Highlights
Various computer vision applications require image quality assessment as a supporting block
The Rectified Linear Unit (ReLU) in the convolutional neural network (CNN) introduces non-linearity to learning image features, and we propose equipping the CNN with a spatial pyramid pooling (SPP) layer[31] to let the network have the ability to deal with images of any arbitrary size
This paper develops an approach to assessing image quality with a deep neural network that comprises a convolutional network, ResNet, and a SPP layer
Summary
Various computer vision applications require image quality assessment as a supporting block. During the process of image/video acquisition, timely monitoring the quality of images enables obtaining images of better quality. In the field of image classification and object detection, the performance relies heavily on the quality of images, and to measure the confidence of classification and detection results, one often needs to have an understanding of image quality. Noise and distortions have a great impact on human’s understanding image content, and so in order to analyze the response of human to compressed images/videos, we often conduct the assessment of image quality for the first step. Objective image quality assessment (IQA) plays an important role in various fields including image acquisition, image super-resolution,[1,2] and image enhancement,[3] etc. IQA is applied in applications such as image reconstruction and image retrieval.[4,5]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
