Unlocking Comics: The AI4VA Dataset for Visual Understanding

Abstract: In the evolving landscape of deep learning, there is a pressing need for more comprehensive datasets capable of training models across multiple modalities. Concurrently, in digital humanities, there is a growing demand to leverage technology for diverse media adaptation and creation, yet limited by sparse datasets due to copyright and stylistic constraints. Addressing this gap, our paper presents a novel dataset comprising Franco-Belgian comics from the 1950s annotated for tasks including depth estimation, semantic segmentation, saliency detection, and character identification. It consists of two distinct and consistent styles and incorporates object concepts and labels taken from natural images. By including such diverse information across styles, this dataset not only holds promise for computational creativity but also offers avenues for the digitization of art and storytelling innovation. This dataset is a crucial component of the AI4VA Workshop Challenges~\url{https://sites.google.com/view/ai4vaeccv2024}, where we specifically explore depth and saliency. Dataset details at \url{https://github.com/IVRL/AI4VA}.

• The paper provides detailed instructions on LNCS formatting, page limits, and LaTeX preparation to standardize submissions.
• It outlines strict guidelines for handling figures, tables, equations, and references to enhance clarity and consistency.
• The guidelines enforce anonymity and adherence to ethical policies, ensuring a fair double-blind review process.

Overview of "Author Guidelines for ECCV Submission"

The document titled "Author Guidelines for ECCV Submission" provides comprehensive instructions for authors preparing submissions to the European Conference on Computer Vision (ECCV). As such, its focus is on ensuring consistency and quality in the presentation of papers for the conference. The guidelines are based on a combination of Springer’s Lecture Notes in Computer Science (LNCS) format and longstanding ECCV traditions.

Key Components of the Guidelines

The document is structured to address several critical areas of paper preparation:

1. Submission Format and Language: All submissions must adhere to the LNCS format, with preparation strongly recommended in \LaTeX. The submission must be in English, with strict adherence to the provided templates to ensure uniformity in presentation.
2. Paper Length and Formatting: The paper should not exceed 14 pages, excluding references. The formatting guidelines are strict, with specific instructions on font usage, section headings, and the overall layout. These constraints are essential to maintain fairness in evaluating the content.
3. Anonymity and Blind Review: Emphasis is placed on maintaining anonymity in submissions to support the double-blind review process. The guidelines provide explicit instructions on how to reference previous work without revealing authorship.
4. Policy Adherence: Authors are reminded of the importance of adhering to ECCV and Springer policies related to confidentiality, dual submissions, and ethical standards. Violations of these policies could lead to rejection.
5. Technical Aspects: Detailed guidance is provided on how to handle figures, tables, equations, and references. There is a stipulation for vector over raster formats for figures to enhance clarity, while references must include DOIs when available.
6. Preserving Standard Conventions: The document advises on standardizing the presentation of technical content, such as the use of "et al." for multi-author works and ensuring equations and figures are numbered and cited properly.

Implications and Future Considerations

These guidelines underscore the importance of standardization in academic publishing. By imposing uniformity in formatting, ECCV ensures that the evaluation of submissions focuses purely on content quality rather than presentation discrepancies. This approach also facilitates a smoother transition to digital publishing and archiving.

As fields like computer vision advance, these guidelines may further evolve to accommodate new types of media and formats. For instance, the incorporation of interactive figures or supplementary datasets might become standardized, reflecting broader trends in open access and reproducibility within the scientific community. Additionally, as automatic formatting tools improve, further integration of such technologies might relieve authors from technical formatting burdens, allowing them to focus more on the academic content.

Conclusion

The "Author Guidelines for ECCV Submission" serve a crucial role in maintaining the integrity and quality of submissions to a leading conference in the field of computer vision. By enforcing a rigorous set of standards, these guidelines help to ensure a level playing field for all submissions, thereby fostering a focus on scientific advancement and innovation. As the landscape of scientific publishing continues to evolve, these guidelines will likely adapt to incorporate and set new standards in research dissemination.