Recently the diversity and multi-functionality of the displays have increased the demands of more advanced image processing methods, such that the quality evaluation and improvement of digital images during the image processing become vital. Since human visual system (HVS) is the ultimate receiver for images, image quality assessment (IQA) algorithms have been proposed to simulate HVS accurately. However, the size of IQA databases limits the evaluation performance of images in the real-world. Fortunately, with the development of deep learning (DL), IQA models based on deep learning have been designed to improve the evaluation ability and generalization performance.

This thesis focuses on the object-oriented and contrast-changed no-reference image quality assessment (NR-IQA) by using handcraft features and DL. It mainly consists of three parts: 1) content-oriented image quality assessment with multi-label SVM classifier; 2) a NR-IQA metric for contrast-changed images via a semi-supervised robust principal component analysis (RPCA); 3) a deep neural network (DNN) based NR-IQA model via object detection. The first two works are explored based on handcraft features and the last one is proposed based on DNN.

In the first part, a content-oriented IQA database is constructed. In particular, the database contains four content types, including landscape, human face, handcrafted scene and hybrid scene. In total, 80 reference images with 20 images for each type of content are involved, and 1600 distorted images with mean opinion scores (MOSs) are generated by using five types and four levels of distortion. Furthermore, to classify these images, especially for the hybrid case, a Support Vector Machine (SVM) based multi-label (ML) classification model is presented.

In the second part, we propose a metric for no-reference quality assessment of contrast-changed images by using a novel semi-supervised RPCA model, which can realize feature selection and denoising simultaneously, guided by the available supervisory information. To select features adaptively, the information-oriented features (e.g. entropy and natural scene statistics) and appearance-oriented features (e.g. colorfulness) are adopted. The proposed model is formulated as a constraint optimization problem, which is further casted to a convex problem and solved via augmented Lagrangian multiplier method.

In the third part, we propose a CNN-based algorithm for NR-IQA via object detection. The network has three parts: an object detector, an image quality prediction network, and a self-correction measurement (SCM) network. First, we detect objects from input image by the object detector. Second, a ResNet-18 network is applied to extract features of the input image and a fully connected (FC) layer is followed to estimate image quality. Third, the features of both the image and its detected objects are extracted by another ResNet-18, then the features of each object are concatenated to the features of the image, respectively. Then, another FC layer is followed to compute the correction value of each object. Finally, the predicted image quality is amended by the SCM values. Experimental results demonstrate that the proposed NR-IQA model has state-of-the-art performance. In addition, cross-databases evaluation indicates the great generalization ability of the proposed model.