Back to search

Aristotle University Of Thessaloniki (AUTH)

A virtual actor prepares: the virtual actor through artificial intelligence tools and acting techniques in the VR Era

Abstract

dc:description

Advances in artificial intelligence (AI), immersive media, and computational modeling are rapidly transforming the performing arts, expanding the expressive possibilities of human actors while introducing new forms of digitally mediated performance. Despite these developments, the systematic integration of acting theory, affective neuroscience, and machine learning for modeling expressive performance remains largely unexplored. The objective of this dissertation is to investigate how the emotional, cognitive, and behavioral dimensions of acting can be measured, structured, and computationally translated into emotionally expressive Virtual Actors (VAs). The concept of the Virtual Actor is proposed as both a theoretical framework and a methodological construct, an embodiment spectrum through which artistic practice, scientific inquiry, and technological innovation converge. To examine the aesthetic, technical, and ethical dimensions of Virtual Actors, the research develops three interconnected directions. First, it introduces the Emotion-Driven Virtual Actor (EDVA) framework, developed through multimodal data collection from professional actors trained in the Stanislavsky system. Neurophysiological signals (EEG and galvanic skin response), audiovisual recordings, facial Action Unit extraction, and self-reported emotional measures were collected and assessed through machine learning pipelines, specifically CPCA+LDA for neural decoding and SVM for acoustic and facial expression classification, to model expressive dynamics and generate emotionally grounded digital embodiments. This framework provides preliminary empirical evidence linking subjective affective states with measurable neural and behavioral correlates. Second, the dissertation develops the Virtual Physical Actor framework for cinematic practice through an AI-based relighting and post-production framework employing Camera-Aligned Material Planes (CAMPs), physically based rendering, and inferred material properties. This methodology enables live-action performances to be embedded within digitally controlled environments, broadening creative flexibility without compromising the actor's expressive integrity. Third, the study proposes actor-centered AI rehearsal systems, including adaptive line-memorization assistants and Meisner-inspired interactive virtual bodies designed to augment, rather than replace, human performers. Within this framework, artificial intelligence functions not as a generator of performance but as an embodied rehearsal collaborator. These frameworks are demonstrated through practice-based realizations, including the proof-of-concept short film Can I Have a Minute?, the Venice Girl project in both parametric and AI-generated variations, and a series of interactive rehearsal prototypes. Across these frameworks, the dissertation analyzes how performance migrates across biological, parametric, and generative embodiment systems, revealing how different computational architectures redistribute expressive control, authorship, and cinematic affordances. This research contributes to affective computing, XR cinema, virtual production, and AI-assisted performance by demonstrating that emotionally intelligent virtual actors can be grounded in human neural and expressive data while remaining artistically controllable. Rather than positioning AI as a substitute for human performers, the dissertation advocates hybrid human–AI collaborations that preserve agency, transparency, and ethical responsibility. The results indicate potential and motivate future investigation of emotionally expressive Virtual Actors across cinematic, immersive, educational, and therapeutic contexts, reaffirming the centrality of human expressive practice within the computational architectures of contemporary digital media. The doctoral research presented in this dissertation was conducted between 2021 and 2026 at the School of Film of the Aristotle University of Thessaloniki, with empirical and practice-based components developed at the USC Institute for Creative Technologies (Summer 2024; IKY-Fulbright Visiting Researcher placement, 2024–2025).

Degree

thesis:*
Grantor dc:publisher
Aristotle University Of Thessaloniki (AUTH)
Year dc:date
2026

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Chavez, Natali-Tereza
  • Τσάβεζ, Ναταλί-Τερέζα

Subjects

dc:subject × 48

Rights

Language dc:language
eng

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/62595

Chain of custody

source
Harvested from
Greek National Archive of PhD Theses
Base URL
phdtheses.ekt.gr/eadd_oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Chavez, Natali-Tereza; Τσάβεζ, Ναταλί-Τερέζα. A virtual actor prepares: the virtual actor through artificial intelligence tools and acting techniques in the VR Era. Aristotle University Of Thessaloniki (AUTH), 2026. http://hdl.handle.net/10442/hedi/62595