Image Editing Of Complex Visual Scene Via Natural Language H/F

CEA; Saclay; Sur place; EUR 40 000 - 60 000; Plein temps

Résumé du poste

A leading research organization is offering an internship focusing on natural language-guided image editing in Saclay. The role involves designing methods for interpreting language to create detailed images and addressing scene complexity. Ideal candidates should be in their final year of studies with strong computer vision and Python skills, especially in deep learning frameworks like PyTorch and TensorFlow.

Qualifications

• Students in their 5th year of studies (M2).
• Strong computer vision and machine learning background.
• Proficiency in Python using deep learning frameworks.

Responsabilités

• Design methods to interpret language for image editing.
• Handle scene complexity for coherence and accuracy.
• Explore NLP and computer vision integration.
• Develop innovative approaches for image generation.

Connaissances

Computer vision skills

Machine learning skills

Python proficiency

Formation

Students in their 5th year of studies (M2)

Outils

PyTorch

TensorFlow

Description du poste

Overview

Image Editing of Complex Visual Scene via Natural Language H/F

Organization

The French Alternative Energies and Atomic Energy Commission (CEA) is a key player in research, development and innovation in four main areas: defence and security; nuclear energy (fission and fusion); technological research for industry; fundamental research in the physical sciences and life sciences. The LIST is part of CEA Tech and focuses on intelligent digital systems. Within LIST, the Laboratory of Vision and Learning for Scene Analysis (LVA) conducts research in computer vision and AI for perception of intelligent and autonomous systems. Based in Saclay (Essonne).

Reference

Reference 2025-37545

Contract

Internship

Job location

Saclay

Subject

This internship focuses on natural language-guided image editing, aiming at generating and modifying complex scenes from verbal descriptions. The candidate will design and implement methods to interpret language for creating or editing detailed images (e.g., crowds, cityscapes, multi-object interactions). Key challenges include managing scene complexity—ensuring coherence and accuracy when multiple objects and relations are involved—and achieving effective multimodal integration between NLP and vision models.

Responsibilities

• Design and implement methods to interpret language for creating or editing detailed images from verbal descriptions.
• Handle scene complexity to maintain coherence and accuracy when multiple objects and relations are present.
• Explore multimodal integration between natural language processing and computer vision models.
• Investigate current methods for natural language-based image generation and editing; develop innovative approaches; demonstrate improvements in accuracy and detail; contribute to academic research through publications and/or patents.

Methods / Means

Computer Vision / ML (deep learning, GenAI…); Python (PyTorch, TensorFlow)

Qualifications

• Students in their 5th year of studies (M2)
• Computer vision skills
• Machine learning skills (deep learning, LLM, VLM, generative AI)
• Python proficiency in a deep learning framework (especially PyTorch or TensorFlow)

Site

Saclay