As a typical information medium, images are widely utilized across various scenarios. Measuring image quality accurately is meaningful for the subsequent usability of images. However, significant variations exist in image types and distortion types in different scenarios. And, acquiring labeled images for each specific scenario is time-consuming and labor-intensive. Consequently, designing cross-domain image quality assessment (IQA) that generalizes across different scenarios remains a substantial challenge. Existing cross-domain IQA methods primarily focus on content relevance while neglecting distortion differences, leading to limited applicability while distortion fluctuates. To address these limitations, a graph-driven domain co-adaptation framework for cross-domain IQA (GDCIQA) is proposed. Firstly, a graph knowledge sharing (GKS) module that constructs graphs via inter-domain distortion relevance has been proposed. GKS employs graph neural networks to update quality-aware features in the source domain by leveraging target-domain representations. Secondly, the proposed co-adaptation learning (CAL) mechanism can enable joint optimization of different modules, which ensures comprehensive sharing of quality-aware and distortion-related information. Finally, a domain adaptation framework has been designed to train models effectively on labeled source images, yielding target-domain-optimized IQA models. Experimental results demonstrate that GDCIQA achieves higher accuracy and stability in cross-domain scenarios. The proposed GKS and CAL can advance cross-domain IQA research.

Most existing image quality assessment methods need to be retrained when dealing with a new type of task. This approach wastes computing resources and time. Therefore, these methods fail to suit the application scenarios that require processing of multi-type image quality assessment tasks. In the human visual system, the eyes of human tend to pay varying degrees of attention to different regions. Inspired by this system, this paper proposes a multi-type image quality assessment method based on multi-region deep feature fusion under meta-learning (MMQA). First, we utilize the differences in the structural information to screen out salient and non-salient regions. Second, a deep multi-stream network is designed to comprehensively consider and fuse different features related to the quality in salient regions, non-salient regions and the entire image. Third, meta-learning is applied to quickly learn and update the parameters of the model when facing new types of images. By summarizing the prior knowledge in the training of one type of task, the model can be quickly fine-tuned for other types of images. The experimental results demonstrate that the proposed method has advantages over the existing methods in generalization and robustness. Furthermore, the proposed method can adapt well to different distortion types and different image types quickly and accurately.

Cross-modal hashing encodes different modalities of multi-modal data into a low-dimensional Hamming space for fast cross-modal retrieval. Most existing cross-modal hashing methods heavily rely on label semantics to boost retrieval performance; however, semantics are expensive to collect in real applications. To mitigate the heavy reliance on semantics, this work proposes a new semi-supervised deep cross-modal hashing method, namely, Graph Convolutional Semi-Supervised Cross-Modal Hashing (GCSCH), which is trained with limited label supervision. The proposed GCSCH first generates pseudo-multi-labels of the unlabeled samples using the simple yet effective idea of consistency regularization and pseudo-labeling. GCSCH designs a fusion network that merges the two modalities and employs Graph Convolutional Network (GCN) to capture semantic information among ground-truth-labeled and pseudo-labeled multi-modal data. Using the idea of knowledge distillation, GCSCH employs a teacher-student learning scheme that can successfully transfer knowledge from the fusion module to the image and text hashing networks. Empirical studies on three multi-modal benchmark datasets demonstrate the superiority of the proposed GCSCH over state-of-the-art cross-modal hashing methods with limited label supervision.

Video hashing learns compact representation by mapping video into low-dimensional Hamming space and has achieved promising performance in large-scale video retrieval. It is challenging to effectively exploit temporal and spatial structure in an unsupervised setting. To fulfill this gap, this paper proposes Contrastive Transformer Hashing (CTH) for effective video retrieval. Specifically, CTH develops a bidirectional transformer autoencoder, based on which visual reconstruction loss is proposed. CTH is more powerful to capture bidirectional correlations among frames than conventional unidirectional models. In addition, CTH devises multi-modality contrastive loss to reveal intrinsic structure among videos. CTH constructs inter-modality and intra-modality triplet sets and proposes multi-modality contrastive loss to exploit inter-modality and intra-modality similarities simultaneously. We perform video retrieval tasks on four benchmark datasets, i.e., UCF101, HMDB51, SVW30, FCVID using the learned compact hash representation, and extensive empirical results demonstrate the proposed CTH outperforms several state-of-the-art video hashing methods.

Image quality assessment (IQA) is a meaningful research topic to meet the increasing demand of high‐quality image. The degradation of image quality will cause changes in image structural information. Meanwhile, human visual system is sensitive to changes in structural information. This finding motivates us to utilise structural information for proposing IQA method which is consistent with human visual perception. Recently, IQA methods are mainly focused on individual image type, e.g. natural image or screen content image (SCI), thus, the authors proposed a novel no‐reference IQA method which can be suitable for both natural image and SCI. The proposed method is based on structural information analysis. For each image, they first obtain the grey‐scale fluctuation maps (GFMs) in four detection directions. After that, the grey‐scale fluctuation direction map (GFD) of certain image can be acquired via its GFMs. Based on the GFMs and GFD, the structural features of each image are extracted, and then collected and transformed to feature vectors. Subsequently, the IQA model is trained by support vector regression. The experimental results on the public databases demonstrate the proposed method can predict image quality accurately for both natural image and SCI, and the performance is competitive with prevalent methods.