Stage Human-Object Interaction benchmarking-Saclay-H/F

CEA; Saclay; Presencial; EUR 12 000 - 16 000; Tempo parcial

Resumo da oferta

Le LIST, laboratoire du CEA, recherche un stagiaire en vision et apprentissage automatique pour étudier l’interprétation des interactions humaines dans les images et vidéos. Travail en perception et apprentissage, avec évaluation et publication prévues.

Le candidat doit être en Master 2 ou équivalent, maîtriser Python et un framework DL (TensorFlow/PyTorch) et être basé à Saclay (Île-de-France).

Qualificações

• Étudiant en fin de cursus (M2 ou gap year) avec des bases solides en vision par ordinateur et ML.
• Compétences en Python et frameworks DL (TensorFlow ou PyTorch) exigées.
• Connaissances en développement et expérimentation de modèles DL en Python.

Responsabilidades

• Réaliser une revue de l’état de l’art sur les bases de données et analyser les biais.
• Proposer un pipeline semi-automatique pour corriger ces biais.
• Identifier les biais et lacunes des métriques utilisées dans l’état de l’art.
• Proposer un nouveau benchmark couvrant divers domaines d’application.
• Évaluer les approches actuelles sur ce benchmark.
• Rédiger une publication sur ce benchmark.

Conhecimentos

Computer vision

Machine learning (DL)

Python (TensorFlow/PyTorch)

Formação académica

Bac+5 - Master 2

Ferramentas

TensorFlow

PyTorch

Descrição da oferta de emprego

Domaine

Contrat

Stage

The interpretation of human interactions in images or videos has significantly improved with the emergence of Large Language Models (LLMs) and Vision-Language Models (VLMs). However, these large models, whether used directly or distilled into specialized models, still have significant limitations, particularly in accurately attributing interactions to the correct person in dense scenes and discriminating actions in the presence of objects.

Evaluation protocols and databases for this task do not always accurately reflect the true capabilities of the methods due to issues such as annotation imprecision or overly rigid semantic metrics. This internship tackles this problem.

Context

The interpretation of human interactions in images or videos has significantly improved with the emergence of Large Language Models (LLMs) and Vision-Language Models (VLMs). However, these large models, whether used directly or distilled into specialized models, still have significant limitations, particularly in accurately attributing interactions to the correct person in dense scenes and discriminating actions in the presence of objects.

Evaluation protocols and databases for this task do not always accurately reflect the true capabilities of the methods due to issues such as annotation imprecision or overly rigid semantic metrics.

What do we expect from you ?

To address these problems, the internship will focus on the following objectives:

• Conduct a state-of-the-art review of existing databases and analyze their biases (e.g., precision of detection boxes).
• Propose a semi-automatic pipeline for correcting these biases.
• Identify the biases and gaps in the metrics commonly used in the state-of-the-art.
• Propose a new benchmark, addressing various application domains.
• Evaluate the main state-of-the-art approaches on this benchmark.
• Write a publication about this benchmark.

Cea List

AI, Deep Neural Network, Computer Vision, Human behavior analysis

Profile

• Students in their 4th or 5th year of studies (M1, M2 or gap year)
• Computer vision skills
• Machine learning skills (deep learning, perception models, generative AI…)
• Python proficiency in a deep learning framework (especially TensorFlow or PyTorch)

Site

Saclay

Saclay

Bac+5 - Master 2

Oui

01/02/2027

Référence

2026-41776

Based in Saclay (Essonne), the LIST is one of the two institutes of CEA Tech, the Technological Research Division of the CEA. Dedicated to intelligent digital systems, its mission is to carry out technological developments of excellence on behalf of industrial partners, in order to create value.

Within the LIST, the Laboratory of Vision and Learning for Scene Analysis (LVA) conducts its research in the field of computer vision and artificial intelligence for the perception of intelligent and autonomous systems. The laboratory's research themes include visual recognition, behavior and activity analysis, large-scale automatic annotation, and perception and decision models.