PhD : "« Making Comics Accessible to People with Visual Impairments: Design and Experimental Evaluation of Immersive AI-Generated Audio Narratives » (BD-Access)" M/F

Job Description

Thesis Subject

Mission :
Background: To ensure equitable access to essential products and services for people with disabilities, the European Union adopted the European Accessibility Act in 2019. This directive requires, among other things, that published e-books be available in a format accessible to people who cannot read visually. However, comic books, due to their visual and narrative complexity, have been granted a temporary exemption by ARCOM. In France, accessible comic books in audio format remain rare and rely on manual adaptations, requiring labor-intensive and costly editorial and artistic work. Although these versions offer a rich listening experience, the cost and complexity of this adaptation limit their number compared to the thousands of comic books published each year.
The BD-Access thesis is part of the ANR EnACA (ENhancing Accessibility in Comic Albums) project, which aims to bridge this accessibility gap by leveraging advances in the fields of artificial intelligence and cognitive science. Converting a visual comic book into an audio narrative involves more than simply verbally reproducing textual information and describing the associated images; it requires constructing an auditory narrative from a multimedia object. It is necessary to robustly extract the various constituent elements of the comic book (detection of panels and speech bubbles, text recognition, segmentation, and character identification), and then to structure this information in order to produce a coherent narrative that respects the narrative intent and the progression of the story. In this context, approaches based on artificial intelligence offer concrete possibilities for automating the analysis of comic book pages, linking elements together, and generating narration scripts. The EnACA Project aims to develop a toolchain ranging from the automatic analysis of comic books to the generation of audio narrations, while integrating these developments with cognitive evaluation among visually impaired audiences and incorporating them into reading standards and technologies. The project employs an interdisciplinary approach based on collaboration among specialists in text and image analysis, natural language processing, large language models, vision-language models, deep learning, and cognitive psychology and ergonomics. The EnACA project consortium includes the Computer Science, Image, and Interaction Laboratory (L3I/EA2118, University La Rochelle), the Fundamental Computer Science Laboratory of Orléans (LIFO/EA 4022, University of Orléans), the Cognition, Languages, Language, and Ergonomics Laboratory (CLLE/UMR 5263, University Toulouse Jean Jaurès), Comix AI, the European Digital Reading Lab, and the Valentin Haüy Association.
The goal of the BD-Access thesis is to draw on theoretical and methodological contributions from psychology and cognitive ergonomics to optimize the comprehension of immersive audio descriptions and the resulting reading experience—particularly in terms of immersion in the story—among adults with visual impairments. Several scientific challenges regarding “what and how to convey” must be addressed:
- the selection and organization of the information to be rendered, since this involves combining information from text and images while making narrative choices: the adopted point of view, the temporal sequence of descriptions, the prioritization of information, and the management of continuity, in order to reduce ambiguities from one panel to the next.
- Choices regarding specific modes of representation (e.g., the voices of characters and the narrator), prosody, and sound design to convey the functions typically provided by layout and graphic cues.
- Job Responsibilities:
Literature review on the following topics: cognitive models of narrative and comic book comprehension, differences between oral and written comprehension, conversion of multimedia content into audio formats (e.g., audio description), accessibility in general and for comic books in particular, immersion, and user experience
Detailed analysis of the accessibility needs of visually impaired users regarding comic books and development of fundamental design principles that enhance their experience. The design will focus on three key dimensions of audio storytelling: cognitive accessibility, ease of interaction, and acceptability.
Experiments - User testing: comparison of the impact of different design principles through controlled user studies, evaluating variables such as story comprehension, subjective mental effort, navigation behaviors, and overall satisfaction. The study will focus on two distinct user groups: (1) users who have been blind since birth (blind since birth or early childhood). (2) users who have lost their sight later in life and those with low vision (those who have lost their sight later in life).
Evaluation of the final accessibility solutions by analyzing their impact on the same variables as before. A key aspect of this evaluation is comparing two user groups: sighted readers who view comics visually and users who are blind or visually impaired. This comparison is crucial for determining whether accessibility solutions provide an experience equivalent to visual reading.
Promotion and dissemination of research findings (oral presentations at international conferences, publications in English, etc.).
Active participation in working meetings of the EnACA and BD-Access project teams, as well as in research seminars organized by the PLC and CLLE teams.
Desired Candidate Profile:
o Degree: Master's degree in cognitive psychology, cognitive ergonomics, or cognitive science
o Required Skills:
- Expertise in cognitive and/or ergonomic psychology, or cognitive science
- Very good knowledge of the cognitive processes involved in language comprehension, multimedia information processing, immersion, and multimodality
- Very good knowledge of user experience evaluation models and techniques
- Strong skills in experimental methodology, usage analysis, and statistical analysis.
- Basic knowledge of AI
- Ability to work in a multidisciplinary collaborative environment (leading meetings, planning, etc.)
- Strong written and oral scientific communication skills
- Very good professional English
- Proficiency in the principles of open science.
- Personal qualities: rigor, motivation, autonomy, proactivity, creativity, perseverance, and adaptability

Your Work Environment

The CLLE Laboratory (Cognition, Languages, Language, Ergonomics) is a CNRS joint research unit (UMR) established in January 2007 and primarily based at the University of Toulouse Jean Jaurès.
The laboratory is part of the CNRS Institute of Humanities and Social Sciences and is primarily affiliated with CNRS Section 34 (Language Sciences) and secondarily with CNRS Section 26 (Cognition, Brain, Behavior) of the CNRS Institute of Biological Sciences. CLLE is therefore a multidisciplinary research unit within the field of Cognitive Sciences: the work conducted there covers the fields—to a greater or lesser extent—of linguistics, psychology, computer science, philosophy, education, and neuroscience.
The doctoral student's work will be part of the Language and Cognitive Processes team, whose research focuses on the interaction between human cognition and language—understood as a cognitive entity, a means of communication, an object of learning and lifelong development, and a sociocultural phenomenon.

Compensation and benefits

Compensation

2300 € gross monthly

Annual leave and RTT

44 jours

Remote Working practice and compensation

Pratique et indemnisation du TT

Transport

Prise en charge à 75% du coût et forfait mobilité durable jusqu’à 300€

About the offer

About the CNRS

The CNRS is a major player in fundamental research on a global scale. The CNRS is the only French organization active in all scientific fields. Its unique position as a multi-specialist allows it to bring together different disciplines to address the most important challenges of the contemporary world, in connection with the actors of change.

CNRS

The research professions