Learning-based Underwater Image Enhancement (UIE) methods have made significant progress. Limited by the manual label selection process, the limited quantity and outdated label quality of UIE datasets have severely hindered the development of UIE society. The urgent demand for more and better paired training samples motivates us to propose Ensemble-Select (E-Select), a strategy that can serve as an alternative to fully manual annotation and can enable continuous expansion of dataset size and optimization of label quality. However, expanding size will encounter new images, optimizing labels will encounter new algorithms and the proposed strategy is required to maintain strong generalization in both scenarios, which overwhelms many IQA methods. This work improves generalization in two ways. First, three primary influencing factors and their interrelationships in quality assessment are systematically analyzed. Specifically, we first explore the interactive relationship between content and distortion perception, and further investigate the guiding value of aesthetic-aware features in image quality perception. Then, the proposed Distortion-Content Interaction Module (DCIM) enables the network to focus on perceptually important distortion features guided by content. Second, we investigate a multi-perspective quality evaluation framework based on the ensemble learning paradigm. Building upon the availability of numerous outstanding IQA works, we initially demonstrate their distinct excel regions and evaluation biases. Subsequently, we explore the ensemble of their results through the proposed Aesthetic-Guided Quality Regression module (AGQR), which generates dynamic quality regression layers and derives image-specific quality perception rules based on aesthetic features. We then construct the first expandable, updatable UIE dataset with the help of E-Select. We collect over 50k real underwater image pairs with optimal labels, covering diverse scenes and varied degradation characteristics. Unlike other datasets, our dataset can consistently expand the number of paired samples and maintain optimal labeling without requiring extensive human labor. Experiments show the facilitating effect of the newly constructed dataset on UIE and the SOTA performance of E-Select. Codes and datasets are available at URL.

Underwater image enhancement (UIE) is crucial for robust marine exploration, yet existing methods prioritize perceptual quality while overlooking irreversible semantic corruption that impairs downstream tasks. Unlike terrestrial images, underwater semantics exhibit layer-specific degradations: shallow features suffer from color shifts and edge erosion, while deep features face semantic ambiguity. These distortions entangle with semantic content across feature hierarchies, where direct enhancement amplifies interference in downstream tasks. Even if distortions are removed, the damaged semantic structures cannot be fully recovered, making it imperative to further enhance corrupted content. To address these challenges, we propose a task-driven UIE framework that redefines enhancement as machine-interpretable semantic recovery rather than mere distortion removal. First, we introduce a multi-scale underwater distortion-aware generator to perceive distortions across feature levels and provide a prior for distortion removal. Second, leveraging this prior and the absence of clean underwater references, we propose a stable self-supervised disentanglement strategy to explicitly separate distortions from corrupted content through CLIP-based semantic constraints and identity consistency. Finally, to compensate for the irreversible semantic loss, we design a task-aware hierarchical enhancement module that refines shallow details via spatial-frequency fusion and strengthens deep semantics through multi-scale context aggregation, aligning results with machine vision requirements. Extensive experiments on segmentation, detection, and saliency tasks demonstrate the superiority of our method in restoring machine-friendly semantics from degraded underwater images. Our code is available at https://github.com/gemyumeng/HSRUIE.

With the rapid global growth of digital fashion, virtual clothing has recently become a hot research topic. However, there is still little literature exploring the mutual influence between virtual clothing design features and consumer purchase intentions. This study delves into the dynamic relationship between virtual clothing design attributes and consumer purchase tendencies, seeking to foster a balanced development of the virtual clothing industry. Following a literature review, this study formulated a theoretical model that includes seven key latent factors: Interactivity, Aesthetic, Creativity, Vividness, Satisfaction, Flow experience, and Purchase Intention. Anchored in the Stimulus-Organism-Response (SOR) paradigm, the study conducted offline trials where respondents tried on virtual clothing using AR as a stimulus, carefully examining data from 295 respondents. The analysis shows that while Interactivity and Aesthetics significantly increase the likelihood of consumer purchases, Creativity and Vividness lack a direct and substantial impact on this intent. Satisfaction and Flow experience are important mediating variables that profoundly influence the purchasing decisions for virtual clothing. The findings of this study provide new insights into virtual clothing design and marketing strategies, emphasizing the importance of optimizing user interaction and Aesthetic experiences in enhancing consumer purchase intentions.

Due to the complex underwater imaging process, underwater images contain a variety of unique distortions. While existing underwater image quality assessment (UIQA) methods have made progress by highlighting these distortions, they overlook the fact that image content also affects how distortions are perceived, as different content exhibits varying sensitivities to different types of distortions. Both the characteristics of the content itself and the properties of the distortions determine the quality of underwater images. Additionally, the intertwined nature of content and distortion features in underwater images complicates the accurate extraction of both. In this paper, we address these issues by comprehensively accounting for both content and distortion information and explicitly disentangling underwater image features into content and distortion components. To achieve this, we introduce a dynamic content-distortion guiding and feature disentanglement network (DysenNet), composed of three main components: the feature disentanglement sub-network (FDN), the dynamic content guidance module (DCM), and the dynamic distortion guidance module (DDM). Specifically, the FDN disentangles underwater features into content and distortion elements, allowing us to more clearly measure their respective contributions to image quality. The DCM generates dynamic multi-scale convolutional kernels tailored to the unique content of each image, enabling content-adaptive feature extraction for quality perception. The DDM, on the other hand, addresses both global and local underwater distortions by identifying distortion cues from both channel and spatial perspectives, focusing on regions and channels with severe degradation. Extensive experiments on UIQA datasets demonstrate the state-of-the-art performance of the proposed method.

Underwater image enhancement (UIE) is a meaningful but challenging task, and many learning-based UIE methods have been proposed in recent years. Although much progress has been made, these methods still have two issues: (1) There exists a significant region-wise quality difference in a single underwater image due to the underwater imaging process, especially in regions with different scene depths. However, existing methods neglect this internal characteristic of underwater images, resulting in inferior performance; (2) Due to the uniqueness of the acquisition approach, underwater image acquisition tools usually capture multiple images in the same or similar scenes. Thus, the underwater images to be enhanced in practical usage are highly correlated. However, when processing a single image, existing methods do not consider the rich external information provided by the related images. There is still room for improvement in their performance. Motivated by these two aspects, we propose a novel internal-external representation learning (UIERL) network to better perform UIE tasks with internal and external information, simultaneously. In the internal representation learning stage, a new depth-based region feature guidance network is designed, including a region segmentation module based on scene depth to sense regions with different quality levels, followed by a region-wise space encoder module. With performing region-wise feature learning for regions with different quality separately, the network provides an effective guidance for global features and thus guides intra-image differentiated enhancement. In the external representation learning stage, we first propose an external information extraction network to mine the rich external information in the related images. Then, internal and external features interact with each other via the proposed external-assist-internal module (external features are updated with the help of internal features) and internal-assist-external module (internal features are updated with the help of external features). In this way, our UIERL fully explores the rich internal and external information to better enhance a single image. All results show that our method can achieve state-of-the-art performance on five benchmarks.

Understanding the values that collaborators bring to a collaboration is important for the design of new systems. In collaborative systems understanding differing values could help design solutions to mitigate conflicts and more effectively coordinate collaboration. We review prior studies of Commons-Based Peer Production (CBPP) identifying four common value dimensions previously noted as present in CBPP: usage value, social value, ideological value, and monetary value. We use this synthetic framework to analyze a dataset of 32 interviews with contributors to Wikimedia Commons and editors of Wikipedia who use Commons resources. Our analysis supports the prior values categories while expanding how some dimensions are expressed by participants. We also highlight four additional value dimensions that were not previously identified in CBPP: cultural heritage value, rarity value, aesthetic value, and administrative value. We discuss the implications of our findings for the design of collaborative systems.

Continuous blood glucose monitoring (CGM) is a bio-sensing technology designed to help individuals with diabetes understand and manage their blood sugar levels. Better control and management have the promise of extending and improving the quality of life with diabetes. For teenagers with Type 1 Diabetes (T1D), CGM has the potential to allow for a more collaborative T1D management and control experience by sharing real-time blood glucose data with caregivers. To understand how well CGM designs are living up to this promise and to identify potential challenges, we used the Asynchronous Remote Communities (ARC) with 16 teenagers with T1D and their informal caregivers over a 6 month period to investigate families' lived experiences with CGM and its data sharing function. We found three challenges in caregiver-teenager collaboration; lack of visibility into behaviors behind data, difficulties balancing teenage daily life with T1D control, and technical limitations of current CGM systems. We also found three issues that hinder informal caregiving coordination; unclear task assignment, difficulties in maintaining standards of care, and a lack of informal caregiver rotation. We provide a number of design implications based on these findings to better facilitate home care for teenagers' T1D.

The study of work practices across two or more collaborative platforms is relatively rare. Participants and researchers have to be competent, or even experts, in both just to begin to make sense of what is happening. With the growing popularity of peer-production systems, the integration of resources across various platforms is more and more common. The framework of stitching is one analytical stance that has been used to describe the cross-platform work to build and highlight informational and social networks. Through a qualitative study with 32 participants who have different foci on Wikipedia and Wikimedia Commons, we reveal their practices in three essential stitching processes, production, curation, and dynamic integration. We highlight how their practices enact stitching and extend the conceptual framework to explain barriers that inhibit effective stitching across platforms. We further discuss implications for research in cross-platform work and design to facilitate Wikipedia-Commons collaboration.

Do models of collaboration among contributors of Wikipedia generalize beyond the larger, western editions of the encyclopedia? In this study, we expanded upon the known collaborative mechanisms on the English Wikipedia and demonstrated that the collaboration model is best captured through the interplay of these mechanisms. We annotated talk page conversations for types of power plays or vies for control over edits that are made to articles, to understand how policy and power play mechanisms in editors' discussions account for behavior in English (EN), Farsi (FA), and Chinese (ZH) language editions of Wikipedia. Our findings show that the same power plays used in EN exist in both FA and ZH but the frequency of their usage differs across the editions. These variations suggest that editors in different language communities value contrasting types of policies to compete for power while discussing and editing articles. Our study contributes to a deeper understanding of how collaboration models developed from a western perspective translate to non-western languages.

Journal writing is an important and common reflective practice in education. Students' reflection journals also offer a rich source of data for formative assessment. However, the analysis of the textual reflections in class of large size presents challenges. Automatic analysis of students' reflective writing holds great promise for providing adaptive real time support for students. This paper proposes a method based on topic modeling techniques for the task of themes exploration and reflection grade prediction. We evaluated this method on a sample of journal writings from pre-service teachers. The topic modeling method was able to discover the important themes and patterns emerged in students' reflection journals. Weekly topic relevance and word count were identified as important indicators of their journal grades. Based on the patterns discovered by topic modeling, prediction models were developed to automate the assessing and grading of reflection journals. The findings indicate the potential of topic modeling in serving as an analytic tool for teachers to explore and assess students' reflective thoughts in written journals.