HomeChair: Josiane B. Zerubia, INRIA, France
Home
Sonia Bouzidi, INRIA (France)
Jean-Paul Berroir, INRIA (France)
Isabelle L Herlin, INRIA (France)
This paper addresses vegetation monitoring in European agricultural areas using Earth Observation satellites. Due to the small size of typical European fields, two complementary sensors are used, SPOT and NOAA-AVHRR, bringing the spatial and the temporal information respectively. A subpixel analysis of NOAA data using one SPOT image is performed to characterize fields with high spatial and temporal resolutions. To be used in an operational context, the method must have realistic data requirements. We define an operational scenario making use of only one SPOT image per site and a one year NOAA sequence,covering a large part of Europe. We first proceed to an unsupervised segmentation of the SPOTimage; the NOAA data analysis on test sites provides the temporal evolution of vegetation; then, identification of fields is performed by minimizing a cost function measuring the similarity between the global reflectance observed on NOAA pixels and the reflectance computed from corresponding regions at SPOT resolution.
ic982440.pdf (From Postscript)
TOP
Nicolas A Tsapatsoulis, National Technical University of Athens (Greece)
Nikolaos D Doulamis, National Technical University of Athens (Greece)
Anastasios D Doulamis, National Technical University of Athens (Greece)
Stefanos D Kollias, National Technical University of Athens (Greece)
A complete face recognition system is proposed in this paper by introducing theconcepts of foreground objects, which are currently used during the MPEG-4standardization phase, to human identification. The system automatically detectsand extracts the human face from the background, even it is not uniform, based ona combination of a retrainable neural network structure and the morphological sizedistribution technique. In order to combine face images of high quality and lowcomputational complexity, the recognition stage is performed in compressed domain.Thus, in contrast to existing recognition schemes, the face images are available intheir original quality and not only in their transformed representation.
ic982487.pdf (From Postscript) TOP
Peter Morguet, Munich University of Technology (Germany)
Manfred Lang, Munich University of Technology (Germany)
Finding and identifying characteristic or meaningful image sequences in a continuous video stream is a challenging task with many applications. This paper presents a new and efficient approach to these temporal segmentation and classification problems based on Hidden Markov Models (HMMs). The basic principle consists in continuously observing the output scores of the HMMs at every time step. Peaks, which appear in the individual HMM output scores, allowto determine in an integral way which image sequence occured at what time. The application of our method to the spotting of connected dynamic hand gestures provided excellent recognition results and a high temporal accuracy.
ic981185.pdf (From Postscript) TOP
Miriam Zacksenhouse, Technion - Institute of Technology (Israel)
Gil Abramovich, Technion - Institute of Technology (Israel)
Gad Hetsroni, Technion - Institute of Technology (Israel)
The distance between texture primitives is of major interest in characterizing texture images. This is especially natural when the texture primitives are elongated structures aligned in parallel to a common main axis, and the distance is measured along the perpendicular axis. Such images arise, for example, in flow visualization studies, where the elongated structures are low-speed streaks. A point process based texture generation model is developed for the one-dimensional texture along lines perpendicular to the streaks. The point process models the location of the edges of the streaks, and using edge detection techniques, its probability density function (pdf) can be estimated by the histogram of the distances between the edges. It is shown that for the studied images the resulting histogram is wide (coefficient of variation greater than half), and demonstrated that in this case, previousely suggested auto-correlation based methods are not adequate.
ic981502.pdf (From Postscript) TOP
Giuseppe Boccignone, Università di Salerno (Italy)
Angelo Chianese, Università di Napoli Federico II (Italy)
Antonio Picariello, Università di Napoli Federico II (Italy)
In this paper we present a method for the detection of microcalcifications in digital mammographic images. Our approach is based on the wavelet transform, but differently from other techniques proposed in the literature, the detection is directly accomplished into the wavelet domain and no inverse transform is required. After a preliminary de-noising pass, microcalcifications are separated from background tissue by exploiting information gained through evaluation of Renyi's entropy at the different decomposition levels of the wavelet space. Experimental results achieved on the standard Nijmegen data set are shown and discussed.
ic981388.pdf (From Postscript) TOP
Vijay Bhatnagar, Wright State University (U.S.A.)
Arnab K. Shaw, Wright State University (U.S.A.)
Rob W Williams, Wright Patterson AFB (U.S.A.)
A new algorithm is presented for Automatic Target Recognition (ATR) where the templates are obtained via Singular Value Decomposition (SVD) of High Range Resolution (HRR) profiles. SVD analysis of a large class of HRR data reveals that the Range-space eigenvectors corresponding to the largest singular value accounts for more than 90% of target energy. Hence, it is proposed that the Range-space eigen-vectors be used as templates for classification. The effectiveness of data normalization and Gaussianization of profile data for improved classification performance is also studied. With extensive simulation studies it is shown that the proposed Eigen-template based ATR approach provides consistent superior performance with recognition rate reaching 99.5% for the four class XPATCH database.
ic982570.pdf (Scanned) TOP
Ara V Nefian, Georgia Institute of Technology (U.S.A.)
Monson H Hayes III, Georgia Institute of Technology (U.S.A.)
The work presented in this paper focuses on the use of Hidden Markov Models for face recognition. A new method based on the extraction of 2D-DCT feature vectors is described, and the recognition results are compared with other face recognition approaches. The method introduced in this paper reduces significantly the computational complexity of previous HMM-based face recognition system, while preserving the same recognition rate.
ic981789.pdf (From Postscript) TOP
Mounir Sayadi, ESSTT (France)
Veronique Buzenac-Settineri, LESTER-UBS (France)
Mohamed Najim, ESI-ENSERB (France)
In this work, we take into account the non gaussian properties of textures and we propose a new approach for their characterization based on bidimensionnal adaptive modelization using high order statistics. The 2D-OLRIV (Bidimensionnal Overdetermined Lattice Recursive Instrumental Variable) algorithm allows accurate texture estimation. Sets of 2D-AR coefficients obtained from the reflection coefficient of the lattice model are used to characterize the texture. This algorithm has the advantage of yielding non biased estimates of the 2D-AR model even when the textured image is disturbed by gaussian noise. A multilayer neural network deals with these coefficients in order to classify different textures. in order to evaluate the performances of this approach, classification sensitivity is evaluated on a set of eight different textures. this charcterization approach gives very promising results.
ic981994.pdf (Scanned) TOP
Laura S Rogers, Vexcel Corporation (U.S.A.)
Carolyn Johnston, Vexcel Corporation (U.S.A.)
In this paper, we present the application of the Type II Local Discriminant Basis (LDB) technique to feature extraction for land use classification in Synthetic Aperature Radar (SAR) images. Our classification algorithm incorporates spatial information into the decision process by classifying small image blocks, instead of single pixels. A feature vector composed of all the values in the image blocks is large for even small image blocks and, therefore, degrades the performance of many classifiers. The LDB technique greatly compresses the dimensionality of the feature vector by indicating the most discriminant coordinates within the wavelet packet decomposition of an image block.
ic982087.pdf (From Postscript) TOP
Etienne G Huot, INRIA (France)
Isabelle L Herlin, INRIA (France)
Repeat-pass SAR interferometric data are multitemporal and display changes occuring between two acquisitions. As a consequence, phase and correlation images contains meaningful informations usable for cropland monitoring. This paper proposes a statistical model to segment high phasimetric structures. It is expressed in a Markov random field framework by using cooperatively phase and correlation information.
ic982450.pdf (From Postscript) TOP