[OPEN] [IML_CG_CV_094] Automated face mesh registration 

Creating photorealistic characters from digital 3D models of actors or humanoid characters has become a standard feature in movie production. The models are initialized from high-resolution 3D captures with an uncontrolled mesh topology, then registered to a reference topology suitable for character animation. Today this registration task, although non-creative, requires manual intervention from artists

The purpose of the internship is to investigate approaches for automating the registration process, within the restricted scope of facial animation. Machine learning approaches could be considered, leveraging an available database of raw and registered 3D facial scans

Skills : machine learning, 3D geometry, deep learning, Python, fluent English

Keywords : 3D registration, machine learning, deep learning

 

This internship is located in Rennes, France. If interested, please apply at stage.rennes@technicolor.com by sending us your resume and a cover letter with the internship reference in the email subject line.

[OPEN] [IML_CG_CV_VP_074] Diminished Reality 

Augmented, Mixed… and now I hear about Diminished Reality? What are the differences?

Standard Augmented Reality (AR) consists in adding virtual overlays on top of the view of your real environment without real blending between real and virtual worlds. Added objects or information might be as simple as numbers or text notifications. Most of the AR apps could be seen as purely informative.

Mixed Reality (MR) is more exciting and more flexible and it offers a wider range of possible use cases. It goes further than AR by creating a new mixed world where physical and digital entities seamlessly interact together. Ideally, everything must be done to create the perfect illusion that virtual objects are real. The geometry as well as the photometry of both real and virtual objects should be consistent. For example, this means that spatial occlusions, physical contacts (collisions, bounces …), virtual shadows should be rendered.

Okay, then, what is Diminished Reality?

Diminished Reality can be seen as a part of MR. It consists in virtually removing real objects from a 3D scene, for example on user request or after collision with a virtual object. It offers also the possibility to virtually move or shrink a real object.

The internship will consist in the analysis of previous work related to Diminished Reality, the specification of a solution suited to our needs, and the implementation, testing and improvement of a method. In a first time we could assume that a textured 3D model of the real scene is available.

A particular attention will be given to the dissemination aspects: presentation to the project team, demonstrator, technical report.

Skills : 3D geometry, Unity3D/C#, real-time rendering - Excellent communication skills and good team player

Keywords : video processing, computer graphics, , 3D object removal, inpainting, mixed/augmented realit

 

This internship is located in Rennes, France. If interested, please apply at stage.rennes@technicolor.com by sending us your resume and a cover letter with the internship reference in the email subject line.

[CLOSED] [IML_CG_009] Relief mapping shadering 

The emergence of Virtual Reality in the media & entertainment industry questions the possibility of good immersive video experience, and not only game experience. User experience with an HMD actually requires a high level of visual comfort, which makes the rendering of 360° flat video not optimal. Instead, it may be required to render the geometry of the environment, whose view should change with the position of the head. This geometry is especially salient for the user at the depth contour of the scene

Techniques able to render a geometry at those contours without an expensive mesh description relies on specific shaders related to displacement/relief mapping and date from roughly 10 years. Internship scope is to explore them further

The candidate will have to:

• Carefully read & analyse the state of art,

• Prototype on a platform to be determined,

• Investigate the perspectives of those techniques (GPU load, performances, ..) in the frame of the hosting project

It should be noted that this work requires starting from the state of art, and that no legacy code has been developed in the lab. Any ideas proposed by the candidate to illustrate or benchmark the techniques will be welcome. There will be a number of possibilities for adapting the final phase of the internship to side investigations, depending of the progress.

Skills : Shader implementation, realtime rendering, knowledge in 3D Engine (Unity/C# typically or Unreal/C++ engines)

Keywords : computer graphics, shader, GPU, OpenGL

 

This internship is located in Rennes, France. If interested, please apply at stage.rennes@technicolor.com by sending us your resume and a cover letter with the internship reference in the email subject line.

[CLOSED] [IML_CG_HCI_012] Design of a Tangible System for Virtual Object Creation in CGI 

Due to the massive use of CGI content for movies and other immersive experiences, it may be necessary to ease the way content is created. In particular today artists are using pen and mouse as interfaces with the computer.

In this internship we would like to investigate new interfaces, such as tangible interfaces, to ease and speed up the creation of assets. We expect that more natural interfaces, using touch, would be more suitable for this artistic process.

The intern will thus have i) to design and create a prototype as interface (an existing sensitive deformable actuator, such as a knit ball, maybe use as a first step), ii) connect it with a CGI creation tool and associated rendering engine, and iii) validate the system.

Skills : Electronics, Computer Graphics, Virtual Reality, C++ programming

Keywords : Virtual Reality, Tangible Interfaces

 

This internship is located in Rennes, France. If interested, please apply at stage.rennes@technicolor.com by sending us your resume and a cover letter with the internship reference in the email subject line.

[CLOSED] [IML_CG_CV_VP_013] Image style transfer from real scene videos to virtual / mixed scene videos 

Deep learning is demonstrating its applicability to a wide set of computer vision problems and in particular it has shown recently its potential for high level image synthesis and manipulation.

In this context, the objective of the internship is to develop a fast method based on neural networks to transfer style in images of real scenes to synthetic images for applications in augmented reality.

The internship will consist of an analysis of the state of the art in style transfer and deep learning, the specification of a solution suited to our needs, and the implementation, testing and improvement of a method.

Photo : An example of image style transfer

Skills : Specific knowledge in computer vision, good programming skills in C/C++ and Python, Windows, possibly CUDA.

Keywords : Computer vision, deep learning, style transfer, mixed/augmented reality, video processing, real-time rendering.

 

This internship is located in Rennes, France. If interested, please apply at stage.rennes@technicolor.com by sending us your resume and a cover letter with the internship reference in the email subject line.

[CLOSED] [IML_CG_014] Facial Information Transfer 

Facial animation is an increasingly important part of the VFX pipeline. Believable virtual characters need to have plausible facial expressions so that the movie audience can relate to them and be immersed in the storyline. But people are extremely sensitive to even fine-scale nuances of facial animation. The VR emergence also strengthens the need for high-fidelity facial animation to improve embodiment of HMD users. The most commonly used technique to animate characters’ faces consists in using blend shapes. They represent local deformations of the face and are decomposed into single units (using the Facial Action Coding System). They are then combined using blending weights to create facial expressions

The goal of this internship is to explore how FACS blend shapes can be transferred to different faces while preserving as much as possible the intensity of the facial deformation. The internship will involve amongst other exploring the quality of transfer of such single units, and how they should be topologically deformed for a specific morphology to maximize the transfer of visual information. The internship will also involve designing specific protocols for perceptual evaluations and user experiments to evaluate the visual fidelity of transfers.

Skills : Python language recommended; Strong background in mathematics; Interest in User Evaluations and Experimentations; Skills in Maya and/or Blender would be an advantage

Keywords : Facial Animation, Animation Transfer, Character Animation, Perception, User Experimentation

 

This internship is located in Rennes, France. If interested, please apply at stage.rennes@technicolor.com by sending us your resume and a cover letter with the internship reference in the email subject line.

[CLOSED] [IML_CG_015] 3D facial acquisition and animation for VR content 

The purpose of the internship is to design, set-up (hardware) and implement (software) a complete pipeline from 3D facial acquisition for VR contents.

The intern will be included in both the VR team and the face rigging team at Technicolor Rennes. Several tools and software developed by the two teams will be used. The work will consist in: (1) setting-up a camera rig to acquire multiple pictures of a character, (2) use a specific software to build the 3D mesh, (3) script the use of a tool to animate the 3D mesh and (4) include the newly created animated face in a VR demo content.

Skills : Computer graphics, Virtual Reality, Python, Editing/Animation tool (Blender, Maya), Unity, fluent in English, motivated by research

Keywords : Virtual Reality, Facial Animation

 

This internship is located in Rennes, France. If interested, please apply at stage.rennes@technicolor.com by sending us your resume and a cover letter with the internship reference in the email subject line.

[CLOSED] [HOME_CG_CV_VP_040] Model Reconstruction Using Multiple Depth Sensors 

This internship focuses on real-time reconstruction of 3D models (Point Clouds) in the context of Telepresence scenarios. The trainee will design real-time algorithms on GPU in order to reconstruct clean and stable 3D point clouds from a “non-rigid” rig of depth cameras such as Kinect V2 or RealSense. The work will take place in research project focusing on Immersive technologies.

Skills : C/C++, GPU Programming (GLSL, Compute Shaders), OpenGL, [Python, Java Script]

Keywords : Reconstruction, Depth Cameras, Computer Vision, Augmented Reality, Mixed Reality, Telepresence, Kinect V2, RealSense.

 

This internship is located in Rennes, France. If interested, please apply at stage.rennes@technicolor.com by sending us your resume and a cover letter with the internship reference in the email subject line.