Ruxandra Tapu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Ruxandra Tapu has a total of 15 co-authors. Here are the top co-authors by number of shared publications:
Bogdan Mocanu
CAMPUS Research Institute, Faculty of ETTI, National University of Science and Technology “Politehnica” Bucharest, Bucharest, Romania
10 shared publications- Titus Zaharia
Samovar, Télécom SudParis, Institut Polytechnique de Paris, Palaiseau, France
10 shared publications - Bogdan Mocanu
Samovar, Télécom SudParis, Institut Polytechnique de Paris
8 shared publications Titus Zaharia
ARTEMIS Department, Institut Polytechnique de Paris, Tilicom SudParis, Laboratoire SAMOVAR, 9 rue Charles Fourier, 91000 Ivry, France and Telecommunication Department, Faculty of ETTI, University “Politehnica” of Bucharest, Splaiul Independentei 313, 060042, Bucharest, Romania
6 shared publications- Antoine Brimont
Samovar, Télécom SudParis, Institut Polytechnique de Paris, Palaiseau, France
3 shared publications - Bogdan Mocanu
Advanced Research and TEchniques for Multidimensional Imaging Systems Department, Institut Mines-Télécom/Télécom SudParis, UMR CNRS MAP5 8145 and 5157 SAMOVAR, 9 rue Charles Fourier, 91000 Évry, France
3 shared publications
Latest Publications
- Open Access
- 4 October 2026
- Published by Association for Computing Machinery (ACM)
- p. 299-309
NewsSegCap: Efficient Dense Video Captioning for TV News
- 4 October 2026
- Published by Association for Computing Machinery (ACM)
- p. 299-309
Open AccessNewsSegCap: Efficient Dense Video Captioning for TV News
- Open Access
- 12 August 2026
- Published by Springer Nature in Multimedia Tools and Applications
- Vol. 85 (8) , 694
Audio-guided articulatory distillation for multilingual visual speech recognition with large language model decoding
Visual Speech Recognition (VSR) remains challenging in multilingual and low-resource settings due to visual ambiguity, limited annotated data, and weak cross-lingual generalization. This paper introduces an audio-guided distillation framework for multilingual VSR that exploits synchronized audio-visual speech during training while preserving visual-only inference. The proposed architecture consists of an audio-visual teacher that learns articulation-aware representations from aligned acoustic and video streams, and a visual-only student trained to approximate the teacher through representation- and decoder-level distillation. To improve transcription under ambiguous visual evidence, continuous speech representations are projected into the embedding space of a pretrained large language model through a lightweight adaptation module, enabling language-conditioned decoding without full LLM retraining. We further introduce RoVSR-II, an extended Romanian in-the-wild VSR corpus comprising approximately 250 h of audiovisual speech, designed to support evaluation in an underrepresented language. Experiments on mTEDx demonstrate consistent improvements over existing multilingual VSR methods across Latin-script languages in terms of Word Error Rate (WER) and Character Error Rate (CER). Additional evaluation on RoVSR-II shows that the proposed model supports zero-shot transfer to Romanian and substantially reduces both WER and CER after parameter-efficient supervised adaptation. Ablation results further confirm the contribution of LLM decoding, teacher initialization, and audio-guided distillation to visual-only recognition performance. © 2017 Elsevier Inc. All rights reserved.
- 12 August 2026
- Published by Springer Nature in Multimedia Tools and Applications
- Vol. 85 (8) , 694
Open AccessAudio-guided articulatory distillation for multilingual visual speech recognition with large language model decoding
Visual Speech Recognition (VSR) remains challenging in multilingual and low-resource settings due to visual ambiguity, limited annotated data, and weak cross-lingual generalization. This paper introduces an audio-guided distillation framework for multilingual VSR that exploits synchronized audio-visual speech during training while preserving visual-only inference. The proposed architecture consists of an audio-visual teacher that learns articulation-aware representations from aligned acoustic and video streams, and a visual-only student trained to approximate the teacher through representation- and decoder-level distillation. To improve transcription under ambiguous visual evidence, continuous speech representations are projected into the embedding space of a pretrained large language model through a lightweight adaptation module, enabling language-conditioned decoding without full LLM retraining. We further introduce RoVSR-II, an extended Romanian in-the-wild VSR corpus comprising approximately 250 h of audiovisual speech, designed to support evaluation in an underrepresented language. Experiments on mTEDx demonstrate consistent improvements over existing multilingual VSR methods across Latin-script languages in terms of Word Error Rate (WER) and Character Error Rate (CER). Additional evaluation on RoVSR-II shows that the proposed model supports zero-shot transfer to Romanian and substantially reduces both WER and CER after parameter-efficient supervised adaptation. Ablation results further confirm the contribution of LLM decoding, teacher initialization, and audio-guided distillation to visual-only recognition performance. © 2017 Elsevier Inc. All rights reserved.
- 3 August 2026
- Published by Springer Nature
- p. 589-604
A Multi-modal BLIP-2 Approach for Video Captioning
- 3 August 2026
- Published by Springer Nature
- p. 589-604
A Multi-modal BLIP-2 Approach for Video Captioning
- Open Access
- 24 July 2026
- Published by Elsevier in Knowledge-Based Systems
- Vol. 351
AI-Based semantic video indexing: a unified survey from segmentation to retrieval
- 24 July 2026
- Published by Elsevier in Knowledge-Based Systems
- Vol. 351
Open AccessAI-Based semantic video indexing: a unified survey from segmentation to retrieval
- Open Access
- 24 June 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Multimedia Computing, Communications, and Applications
- Vol. 22 (7) , 1-35
A Survey on Video Captioning in the Era of Large Language Models
Recent technological advancements have led to the widespread presence of video in our daily lives, particularly through social media, where it has become the dominant medium of communication. As a result, the ability to analyze and understand videos has become essential for managing such a massive volume of content. Within this context, video captioning (VC) is one of the most effective ways to achieve this understanding. Recent advances in Natural Language Generation, combined with breakthroughs in Computer Vision, has facilitated the development of robust VC models. However, despite these advancements, selecting the right model for the right application remains challenging due to limited evaluation methods, and significant limitations still hinder the widespread adoption of these models. This survey aims to address these issues by proposing a comprehensive overview of the VC models landscape, along with a comparative experimental study, carried out on two benchmark datasets and involving eight representative, high quality models.
- 24 June 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Multimedia Computing, Communications, and Applications
- Vol. 22 (7) , 1-35
Open AccessA Survey on Video Captioning in the Era of Large Language Models
Recent technological advancements have led to the widespread presence of video in our daily lives, particularly through social media, where it has become the dominant medium of communication. As a result, the ability to analyze and understand videos has become essential for managing such a massive volume of content. Within this context, video captioning (VC) is one of the most effective ways to achieve this understanding. Recent advances in Natural Language Generation, combined with breakthroughs in Computer Vision, has facilitated the development of robust VC models. However, despite these advancements, selecting the right model for the right application remains challenging due to limited evaluation methods, and significant limitations still hinder the widespread adoption of these models. This survey aims to address these issues by proposing a comprehensive overview of the VC models landscape, along with a comparative experimental study, carried out on two benchmark datasets and involving eight representative, high quality models.
- 4 June 2026
- Published by Springer Nature
- p. 81-88
Evaluating Speech Emotion Recognition Systems: From Traditional Low-Level Features to Transformer-Based Models
- 4 June 2026
- Published by Springer Nature
- p. 81-88
Evaluating Speech Emotion Recognition Systems: From Traditional Low-Level Features to Transformer-Based Models
- 1 May 2026
- Published by Springer Nature
- p. 290-297
Food Image Recognition: From CNNs to Transformers and Multimodal Learning
- 1 May 2026
- Published by Springer Nature
- p. 290-297
Food Image Recognition: From CNNs to Transformers and Multimodal Learning
- Open Access
- 25 April 2026
- Published by Elsevier in Computer Vision and Image Understanding
- Vol. 268
EARS4SEE: A multimodal audio description system dedicated to blind and visually impaired users
- 25 April 2026
- Published by Elsevier in Computer Vision and Image Understanding
- Vol. 268
Open AccessEARS4SEE: A multimodal audio description system dedicated to blind and visually impaired users
- Open Access
- 27 January 2026
- Published by MDPI AG in Applied Sciences
- Vol. 16 (3) , 1289
A Comparative Study of Emotion Recognition Systems: From Classical Approaches to Multimodal Large Language Models
Emotion recognition in video (ERV) aims to infer human affect from visual, audio, and contextual signals and is increasingly important for interactive and intelligent systems. Over the past decade, ERV has evolved from handcrafted features and task-specific deep learning models toward transformer-based vision–language models and multimodal large language models (MLLMs). This review surveys this evolution, with an emphasis on engineering considerations relevant to real-world deployment. We analyze multimodal fusion strategies, dataset characteristics, and evaluation protocols, highlighting limitations in robustness, bias, and annotation quality under unconstrained conditions. Emerging MLLM-based approaches are examined in terms of performance, reasoning capability, computational cost, and interaction potential. By comparing task-specific models with foundation model approaches, we clarify their respective strengths for resource-constrained versus context-aware applications. Finally, we outline practical research directions toward building robust, efficient, and deployable ERV systems for applied scenarios such as assistive technologies and human–AI interaction.
- 27 January 2026
- Published by MDPI AG in Applied Sciences
- Vol. 16 (3) , 1289
Open AccessA Comparative Study of Emotion Recognition Systems: From Classical Approaches to Multimodal Large Language Models
Emotion recognition in video (ERV) aims to infer human affect from visual, audio, and contextual signals and is increasingly important for interactive and intelligent systems. Over the past decade, ERV has evolved from handcrafted features and task-specific deep learning models toward transformer-based vision–language models and multimodal large language models (MLLMs). This review surveys this evolution, with an emphasis on engineering considerations relevant to real-world deployment. We analyze multimodal fusion strategies, dataset characteristics, and evaluation protocols, highlighting limitations in robustness, bias, and annotation quality under unconstrained conditions. Emerging MLLM-based approaches are examined in terms of performance, reasoning capability, computational cost, and interaction potential. By comparing task-specific models with foundation model approaches, we clarify their respective strengths for resource-constrained versus context-aware applications. Finally, we outline practical research directions toward building robust, efficient, and deployable ERV systems for applied scenarios such as assistive technologies and human–AI interaction.
- 17 January 2026
- Published by Springer Nature
- p. 85-97
Seeing Through Words: A Zero-Shot Multimodal Audio Description System with Foundation Models
- 17 January 2026
- Published by Springer Nature
- p. 85-97
Seeing Through Words: A Zero-Shot Multimodal Audio Description System with Foundation Models