Optimization of media processing workflows with adaptive operator behaviors

In this paper, we present the ARIA media processing workflow architecture that processes, filters, and fuses sensory inputs and actuates responses in real-time. The components of the architecture are programmable and adaptable; i.e. the delay, size, and quality/precision characteristics of the individual operators can be controlled via a number of parameters. Each data object processed by qStream components is subject to transformations based on the parameter values. For instance, the quality of an output data object and the corresponding processing delay and resource usage depend on the values assigned to parameters of the operators in the object flow path. In Candan, Peng, Ryu, Chatha, Mayer (Efficient stream routing in quality- and resource-adaptive flow architectures. In: Workshop on multimedia information systems, 2004), we introduced a class of flow optimization problems that promote creation and delivery of small delay or small resource-usage objects to the actuators in single-sensor, single-actuator workflows. In this paper, we extend our attention to multi-sensor media processing workflow scenarios. The algorithms we present take into account the implicit dependencies between various system parameters, such as resource consumption and object sizes. We experimentally show the effectiveness and efficiency of the algorithms.

How do designers represent to themselves the users' needs?

This paper reports on an ergonomic study carried out during the design of a cutting machine-tool for the composite material in carbody parts casting. During this design process, the users’ needs were inferred by the designers on the basis of their own mental representations of the use of the new device. These representations of the users’ needs, correct or false, play a decisive role in the choice of a solution. The aim of the study is to identify their particularity. Analyzing the design meetings, we have highlighted that users are considered either as subsystems or basic design principles or elements of an imagined scenario. We have shown that these representations are linked to the types of meetings held during the design process. Accordingly, a diversification of the types of meetings should be promoted by the project leader, so that designers extend their points of view of the operators.     

Combining ergonomics, culture and scenario for the design of a cooperation platform

Analyzing the way computer technologies are used is crucial for their development. Such analyses make it possible to evaluate these technologies and enhance their evolution. The present article presents some ideas drawn from the development of a cooperation platform for elementary school children (10–11 years old). On the basis of an obvious ergonomic requirement, we worked on two other dimensions: cultural aspects and the teaching scenario. The goal was to set up observation situations and analyze the conversations produced during those situations, in order to understand what using the platform meant to both the pupils and their teachers.

On Similarity Measures for Multimedia Database Applications

A multimedia database query consists of a set of fuzzy and boolean (or crisp) predicates, constants, variables, and conjunction, disjunction, and negation operators. The fuzzy predicates are evaluated based on different media criteria, such as color, shape, layout, keyword. Since media-based evaluation yields similarity values, results to such a query is defined as an ordered set. Since many multimedia applications require partial matches, query results also include tuples which do not satisfy all predicates. Hence, any fuzzy semantics which extends the boolean semantics of conjunction in a straight forward manner may not be desirable for multimedia databases. In this paper, we focus on the problem of ‘given a multimedia query which consists of multiple fuzzy and crisp predicates, how to provide the user with a meaningful overall ranking.’ More specifically, we study the problem of merging similarity values in queries with multiple fuzzy predicates. We describe the essential multimedia retrieval semantics, compare these with the known approaches, and propose a semantics which captures the retrieval requirements in multimedia databases. Received 13 August 1999 / Revised 13 May 2000 / Accepted in revised form 26 July 2000     

Software assessment using metrics: A comparison across large C++ and Java systems

This paper presents an assessment method to evaluate the quality of object oriented software systems. The assessment method is based on source code abstraction, object–oriented metrics and graphical representation. The metrics used and the underlying model representing the software are presented. The assessment method experiment is part of an industrial research effort with the Bell Canada Quality Engineering and Research Group. It helps evaluators assess the quality and risks associated with software by identifying code fragments presenting unusual characteristics. The assessment method evaluates object–oriented software systems at three levels of granularity: system level, class level and method level. One large C++ and eight Java software systems, for a total of over one million lines of code, are presented as case studies. A critical analysis of the results is presented comparing the systems and the two languages.     

Accelerated multiplicative updates and hierarchical ALS algorithms for nonnegative matrix factorization

Nonnegative matrix factorization (NMF) is a data analysis technique used in a great variety of applications such as text mining, image processing, hyperspectral data analysis, computational biology, and clustering. In this letter, we consider two well-known algorithms designed to solve NMF problems: the multiplicative updates of Lee and Seung and the hierarchical alternating least squares of Cichocki et al. We propose a simple way to significantly accelerate these schemes, based on a careful analysis of the computational cost needed at each iteration, while preserving their convergence properties. This acceleration technique can also be applied to other algorithms, which we illustrate on the projected gradient method of Lin. The efficiency of the accelerated algorithms is empirically demonstrated on image and text data sets and compares favorably with a state-of-the-art alternating nonnegative least squares algorithm.     

Estimating the Natural Illumination Conditions from a Single Outdoor Image

Given a single outdoor image, we present a method for estimating the likely illumination conditions of the scene. In particular, we compute the probability distribution over the sun position and visibility. The method relies on a combination of weak cues that can be extracted from different portions of the image: the sky, the vertical surfaces, the ground, and the convex objects in the image. While no single cue can reliably estimate illumination by itself, each one can reinforce the others to yield a more robust estimate. This is combined with a data-driven prior computed over a dataset of 6 million photos. We present quantitative results on a webcam dataset with annotated sun positions, as well as quantitative and qualitative results on consumer-grade photographs downloaded from Internet. Based on the estimated illumination, we show how to realistically insert synthetic 3-D objects into the scene, and how to transfer appearance across images while keeping the illumination consistent.