Marius Preda

Marius Preda
Institut Mines-Télécom | telecom-sudparis.eu · ARTEMIS: Advanced Research and Techniques for Multidimensional Imaging Systems

PhD

About

314
Publications
24,267
Reads
How we measure 'reads'
A 'read' is counted each time someone views a publication summary (such as the title, abstract, and list of authors), clicks on a figure, or views or downloads the full-text. Learn more
989
Citations
Introduction
Currently working on point cloud compression and augmented reality

Publications

Publications (314)
Preprint
Full-text available
Neural network pruning is a widely used strategy for reducing model storage and computing requirements. It allows to lower the complexity of the network by introducing sparsity in the weights. Because taking advantage of sparse matrices is still challenging, pruning is often performed in a structured way, i.e. removing entire convolution filters in...
Article
This research work aims to provide an AR training system adapted to industry, by addressing key challenges identified during a long-term case study conducted in a boiler-manufacturing factory. The proposed system relies on low-cost visual assets (i.e., text, image, video and predefined auxiliary content) and requires solely a head-mounted display (...
Preprint
Full-text available
In recent years, deep neural networks have known a wide success in various application domains. However, they require important computational and memory resources, which severely hinders their deployment, notably on mobile devices or for real-time applications. Neural networks usually involve a large number of parameters, which correspond to the we...
Conference Paper
Augmented Reality (AR) is an emerging technology offering a great potential in assisting humans in a wide range of industrial processes, from manufacturing to validation and maintenance. However, very few AR solutions have been adopted so far in industrial sectors, mainly because of technical and acceptability issues. This paper has three main cont...
Conference Paper
The adoption of Augmented Reality (AR) in the industry is in early stages, mainly due to technological and organizational limitations. This research work, carried out in a manufacturing factory, aims at providing an effective AR training method for manual assembly, adapted for industrial context. We define the 2W1H (What, Where, How) principle to f...
Preprint
Full-text available
Introducing sparsity in a neural network has been an efficient way to reduce its complexity while keeping its performance almost intact. Most of the time, sparsity is introduced using a three-stage pipeline: 1) train the model to convergence, 2) prune the model according to some criterion, 3) fine-tune the pruned model to recover performance. The l...
Article
In this article, a survey of the point cloud compression (PCC) methods by organizing them with respect to the data structure, coding representation space, and prediction strategies is presented. Two paramount families of approaches reported in the literature--the projection- and octree-based methods--are proven to be efficient for encoding dense an...
Conference Paper
Full-text available
Augmented Reality (AR) is a rapidly growing field in information and communication technologies, drawing increasing numbers of professionals. Higher education institutions, however, are struggling to keep abreast of its development and to train specialists quickly, providing few courses which sufficiently align with the needs of industry. In additi...
Conference Paper
Full-text available
With the rapid development of point cloud acquisition technologies, high-quality human-shape point clouds are more and more used in VR/AR applications and in general in 3D Graphics. To achieve near-realistic quality, such content usually contains an extremely high number of points (over 0.5 million points per 3D object per frame) and associated att...
Conference Paper
Full-text available
This paper outlines the objectives of the working group on developing a model Augmented Reality curriculum for higher education. We motivate the need for the model curriculum by the growing Augmented Reality industry and subsequent demand for trained professionals. While the industry is growing, the educational offers that train the required skills...
Article
Full-text available
3D point cloud is a simple data structure representing both static and dynamic 3D objects. Though providing close-to-reality visualization, point clouds with high density raise massive demand for storage while sparse point clouds used for navigation require high precision. The emergency of compression technologies is mandatory in the form of standa...
Article
After decennia of developing leading-edge 2D video compression technologies, the Moving Picture Expert Group (MPEG) is currently working on the new era of coding for immersive applications, referred to as MPEG-I, where “I” refers to the “Immersive” aspects. It ranges from 360° video with head-mounted displays to free navigation in 3D space with hea...
Conference Paper
Full-text available
In recent years, 3D point clouds have enjoyed a great popularity for representing both static and dynamic 3D objects. When compared to 3D meshes, they offer the advantage of providing a simpler, denser and more close-to-reality representation. However, point clouds always carry a huge amount of data. For a typical example of a point cloud with 0.7...
Article
Consumer- and industry-level 3d sensing devices are becoming more common than ever before, increasing the amount of available 3D point-cloud data. The full geometry and details of a 3D scene can be captured by 3D scans, which are useful in many applications, including virtual reality (VR), 3D video, robotics, and geographic information access. Howe...
Article
The widespread adoption of new 3D sensor and authoring technologies has made it possible to capture 3D scenes and models in real time with decent visual quality. As an example, Microsoft's Kinect and Apple's PrimeSense technology are now being used in a wide variety of interactive 3D mobile applications, including gaming and augmented reality appli...
Article
Due to the increased popularity of augmented and virtual reality experiences, the interest in capturing the real world in multiple dimensions and in presenting it to users in an immersible fashion has never been higher. Distributing such representations enables users to freely navigate in multi-sensory 3D media experiences. Unfortunately, such repr...
Article
Full-text available
Recently, Internet of Things (IoT) drives a large variety of research, development, and new type of markets. All type of devices and sensors will be part of the Internet of Things and will be able to communicate not only plain data, but also audio-visual, olfactory, and haptic media data. In addition, as the devices and sensors getting smarter, it...
Conference Paper
Full-text available
In this paper, we propose a new watermarking method based on Harris interest points and Fourier Clifford Transform. We employed Harris detector to select robust interest points and to generate some non-overlapped circular interest regions. Each region was transformed into Clifford Fourier domain and the watermark was embedded into the Clifford tran...
Chapter
Virtual worlds such as Second Life and those used in 4D Internet/broadcasting services have become increasingly popular. A life-scale virtual-world presentation and the intuitive interaction between users and virtual worlds can provide a more natural and immersive experience for users. MPEG-V specifies the associated information representations to...
Chapter
In an MPEG-V environment, there are two types of devices: actuators and sensors. Actuators are used to generate sensorial effects in the real world or render them. Sensors are used to pick up user inputs or environmental information for adaptation of the effects, or use them as controlling commands for a virtual world. MPEG-V provides standardized...
Chapter
All actuators and sensors used in rendering sensorial effects have their own capabilities or sensing environments. When commanding an actuator to render an effect, or when obtaining information from a sensor for control in a virtual world, the capabilities of the devices, such as the operating range of an actuator or accuracy of a sensor, should be...
Chapter
There are several tools or types which are used in multiple parts of the MPEG-V standard. MPEG-V Part 6 provides definitions and examples for the use of these common tools and types. Most of the classification schemes defined for MPEG-V are also provided in MPEG-V Part 6. A classification scheme is a collection of terms or objects in a certain cate...
Conference Paper
This paper presents work in progress of the European Commission FP7 project BRIDGET "BRIDging the Gap for Enhanced broadcasT". The project is developing innovative technology and the underlying architecture for efficient production of second screen applications for broadcasters and media companies. The project advancements include novel front-end a...
Conference Paper
Dans cet article, nous proposons une méthode pour la reconnaissance de formes sous une plateforme mobile Android. L'objet d'intérêt est présenté par un ensemble de descripteurs calculés à partir de points caractéristiques. Les descripteurs sont comparés avec un ensemble de donnés de références. La fonction d'appariement conçu est optimisée pour fai...
Article
In this paper we present a novel approach for object identification and tracking in large image datasets. Objects of interest are represented by feature points and descriptors extracted and compared to a set of reference data. An optimized matching paradigm is designed to deal with scalable image databases while keeping a good recognition rate in r...
Article
This book is the first to cover the recently developed MPEG-V standard, explaining the fundamentals of each part of the technology and exploring potential applications. Written by experts in the field who were instrumental in the development of the standard, this book goes beyond the scope of the official standard documentation, describing how to u...
Article
This paper presents an approach for tracking natural objects in augmented reality applications. The targets are detected and identified using a markerless approach relying upon the extraction of image salient features and descriptors. The method deals with large image databases using a novel strategy for feature retrieval and pairwise matching. Fur...
Article
In this paper we are presenting several issues, possible solutions and future trends in addressing 3D aspects in Augmented Reality and Augmented Virtuality
Conference Paper
Transmodality is the partitioning of an image into regions that are expected to present a better entropy using different coding schemes, depending on their structural density, at constant bandwidth. In this paper we present the transmodality of video stream. Our contribution is a transmoder module that includes various different optimized video cod...
Article
ISO/IEC JTC1/SC29/WG11 M32309
Article
ISO/IEC JTC1/SC29/WG11 MPEG2011/m32337
Article
ISO/IEC JTC1/SC29/WG11 MPEG2011/m32335
Conference Paper
The Internet of Things (IoT) has captured the attention of industry as the impetus of new growth. For the successful industrial deployment of Internet of Things services, it is important to provide a standardized means of describing “Things.” MPEG (ISO/IEC SC29 WG11) developed several standards related to sensors, actuators, and multimedia platform...
Article
This paper provides an overview of the rationale of the Reconfigurable Media Coding framework developed by MPEG standardization committee to overcome the limits of traditional ways of providing decoder specifications. Such framework is an extension of the Reconfigurable Video coding framework now encompassing also 3D Graphics coding standard. The i...
Article
The main objective of this paper is to introduce the concept of Reconfigurable Graphic Coding and its validation under the form of a Functional Units (FU) library. The heterogeneity of data for 3D graphics objects representation requires the adaptability of the compression schemas to various types of content. While such adaptation can be relatively...
Article
ISO/IEC JTC1/SC29/WG11 MPEG2011/m31418
Conference Paper
Full-text available
This paper presents a natural feature tracking system for object recognition in real-life environments. The system is based on a local keypoint descriptor method optimized and adapted to extract salient regions within the image. Each object in the gallery is characterized by keypoints and corresponding local descriptors. The method first identifies...
Article
ISO/IEC JTC1/SC29/WG11 MPEG2012/ m30576
Conference Paper
In this work we present a novel interaction approach based on a gesture recognition system using a Microsoft Kinect sensor. Gestures are defined and interpreted in order to activate controls on a media device. This natural interface enables intuitive interaction with the multimedia content. The depth sensor observes the scene to detect a request fo...
Article
ISO/IEC JTC1/SC29/WG11 MPEG2013/m29497
Article
ISO/IEC JTC 1/SC 29/WG 11 m29453