Text localization and recognition in images is important for searching information in digital photo archives, video databases and Web sites. However, since text is often printed against a complex background, it is often difficult to detect. In this paper, a robust text localization approach is presented, which can automatically detect horizontally aligned text with different sizes, fonts, colors and languages. First, a wavelet transform is applied to the image and the distribution of high-frequency wavelet coefficients is considered to statistically characterize text and non-text areas. Then, the k-means algorithm is used to classify text areas in the image. The detected text areas undergo a projection analysis in order to refine their localization. Finally, a binary segmented text image is generated, to be used as input to an OCR engine. The detection performance of our approach is demonstrated by presenting experimental results for a set of video frames taken from the MPEG-7 video test set.

People increasingly use videos on the Web for learning, be it for daily tasks in formal or informal educational settings. To enhance this type of learning, scientists are continuously conducting experiments, proposing guidelines, analyzing data, and researching artificial intelligence methods for tool development. In this comprehensive review, we analyze 257 articles (using PRISMA guidelines) on video-based learning from a technological perspective for the period from 2016 to 2021. One of the aims is to identify video characteristics that support learning as explored by previous research. Based on our analysis, we suggest a taxonomy that organizes the video characteristics and contextual aspects into eight categories: (1) audio features, (2) visual features, (3) textual features, (4) instructor behavior, (5) learners’ activities (play, pause, etc.), (6) interactive features (quizzes, etc.), (7) production style, and (8) instructional design. Also, we identify four representative methodological approaches: (1) tool support of video-based learning, (2) controlled experiments, (3) data analysis studies, and (4) design guidelines for learning videos. We find that the most explored characteristics are textual features followed by visual features, learners’ activities, and interactive features. Tools that aid learning through videos frequently utilize text from transcripts, video frames, and images. The learner’s activity is heavily explored through log files in data analysis studies, and interactive features are frequently scrutinized in controlled experiments. As further contributions, we contrast research findings on how video characteristics affect learning effectiveness, report on tasks and technologies used to develop tools, and summarize design guideline trends to produce learning videos. Our findings provide actionable insights for the design of intelligent educational systems that better support video-based learning.

The increasing proliferation of misinformation and its alarming impact have motivated both industry and academia to develop approaches for misinformation detection and fact checking. Recent advances on large language models (LLMs) have shown remarkable performance in various tasks, but their potential in misinformation detection remains relatively underexplored. Most of existing state-of-the-art approaches either do not consider evidence and solely focus on claim related features or assume the evidence is provided. Few approaches consider evidence retrieval as part of the misinformation detection but rely on fine-tuning models. In this paper, we investigate the potential of LLMs for misinformation detection in a zero-shot setting. We incorporate an evidence retrieval component as it is crucial to gather pertinent information from various sources to detect the veracity of claims. To this end, we propose a novel re-ranking approach for multimodal evidence retrieval using both LLMs and large vision-language models (LVLM). The retrieved evidence samples (images and texts) serve as the input for an LVLM-based approach for multimodal fact verification (LVLM4FV). To enable a fair evaluation, we address the issue of incomplete ground truth in an existing evidence retrieval dataset by annotating a more complete set of evidence samples for both image and text retrieval. Our experimental results on two datasets demonstrate the superiority of the proposed approach in both evidence retrieval and fact verification tasks, with a better generalization capability.

We examine the novel task of domain-independent scientific concept extraction from abstracts of scholarly articles and present two contributions. First, we suggest a set of generic scientific concepts that have been identified in a systematic annotation process. This set of concepts is utilised to annotate a corpus of scientific abstracts from 10 domains of Science, Technology and Medicine at the phrasal level in a joint effort with domain experts. The resulting dataset is used in a set of benchmark experiments to (a) provide baseline performance for this task, (b) examine the transferability of concepts between domains. Second, we present a state-of-the-art deep learning baseline. Further, we propose the active learning strategy for an optimal selection of instances from among the various domains in our data. The experimental results show that (1) a substantial agreement is achievable by non-experts after consultation with domain experts, (2) the baseline system achieves a fairly high F1 score, (3) active learning enables us to nearly halve the amount of required training data. Keywords Sequence labelling Information extraction Scientific articles Active learning Scholarly communication Research knowledge graph