Ilias Grinias - Academia.edu (original) (raw)

Papers by Ilias Grinias

ISPRS Journal of Photogrammetry and Remote Sensing, 2016

We present in this article a new method on unsupervised semantic parsing and structure recognitio... more We present in this article a new method on unsupervised semantic parsing and structure recognition in peri-urban areas using satellite images. The automatic "building" and "road" detection is based on regions extracted by an unsupervised segmentation method. We propose a novel segmentation algorithm based on a Markov random field model and we give an extensive data analysis for determining relevant features for the classification problem. The novelty of the segmentation algorithm lies on the classdriven vector data quantization and clustering and the estimation of the likelihoods given the resulting clusters. We have evaluated the reachability of a good classification rate using the Random Forest method. We found that, with a limited number of features, among them some new defined in this article, we can obtain good classification performance. Our main contribution lies again on the data analysis and the estimation of likelihoods. Finally, we propose a new method for completing the road network exploiting its connectivity, and the local and global properties of the road network.

We propose a method for interactive colour image segmentation. The goal is to detect an object fr... more We propose a method for interactive colour image segmentation. The goal is to detect an object from the background, when some markers on object(s) and the background are given. As features only probability distributions of the data are used. At first, all the labelled seeds are independently propagated for obtaining homogeneous connected components for each of them. Then the image is divided in blocks, which are classified according to their probabilistic distance from the classified regions. A topographic surface for each class is obtained, using Bayesian dissimilarities and a min-max criterion. Two algorithms are proposed: a regularized classification based on the topographic surface and incorporating an MRF model, and a priority multi-label flooding algorithm. Segmentation results on the LHI data set are presented.

In this paper we present a video summarization scheme. First, shot detection is performed and the... more In this paper we present a video summarization scheme. First, shot detection is performed and then we extract the key frames under the equality principle. We propose a key frames selection algorithm (Iso-Content MINMAX), which is very flexible on any changes of content descriptors, based on MINMAX optimization formulation. The equality principle provides to the selected key frames the useful property to be equivalent on content video summarization. EP Algorithm P (Descriptors) Function d(x,y) Computation M sub-shots Key Frames Selection Scheme M (number of key frames) M Key Frames

EURASIP Journal on Advances in Signal Processing, 2003

EURASIP Journal on Advances in Signal Processing, 2002

The algorithm presented in this paper is comprised of three main stages: (1) classification of th... more The algorithm presented in this paper is comprised of three main stages: (1) classification of the image sequence and, in the case of a moving camera, parametric motion estimation, (2) change detection having as reference a fixed frame, an appropriately selected frame or a displaced frame, and (3) object localization using local colour features. The image sequence classification is based on statistical tests on the frame difference. The change detection module uses a two-label fast marching algorithm. Finally, the object localization uses a region growing algorithm based on the colour similarity. Video object segmentation results are shown using the COST 211 data set.

IEEE Transactions on Image Processing, 2011

We propose a general purpose image segmentation framework, which involves feature extraction and ... more We propose a general purpose image segmentation framework, which involves feature extraction and classification in feature space, followed by flooding and merging in spatial domain. Region growing is based on the computed local measurements and distances from the distribution of features describing the different classes. Using the properties of the label dependent distances spatial coherence is ensured, since the image features are described globally. The distribution of the features for the different classes are obtained by block-wise unsupervised clustering based on the construction of the minimum spanning tree of the blocks' grid using the Mallows distance and the equipartition of the resulting tree. The final clustering is obtained by using the k-centroids algorithm. With high probability and under topological constraints, connected components of the maximum likelihood classification map are used to compute a map of initially labelled pixels. An efficient flooding algorithm is introduced, namely, Priority Multi-Class Flooding Algorithm (PMCFA), that assign pixels to labels using Bayesian dissimilarity criteria. A new region merging method, which incorporates boundary information, is introduced for obtaining the final segmentation map. Therefore, the merging stage is based on region features and edge localization. Segmentation results on the Berkeley benchmark data set demonstrate the effectiveness of the proposed methods.

We present an unsupervised, automatic human motion analysis and action recognition scheme tested ... more We present an unsupervised, automatic human motion analysis and action recognition scheme tested on athletics videos. First, four major human points are recognized and tracked using human silhouettes that are computed by a robust camera estimation and object localization method. Statistical analysis of the tracking points motion obtains a temporal segmentation on running and jump stage. The method is tested on athletics videos of pole vault, high jump, triple jump and long jump recognizing them using robust and independent from the camera motion and the athlete performance features. The experimental results indicate the good performance of the proposed scheme, even in sequences with complicated content and motion.

IEEE Transactions on Image Processing, 2011

We describe a user-guided system for the segmentationof an image sequence. A first segmentation w... more We describe a user-guided system for the segmentationof an image sequence. A first segmentation which involvestwo images of the sequence is presented, as wellas the tracking of its result during a number of frames.The segmentation algorithm is a Region Growing algorithm.The main segmentation feature is the motionof the objects presented in the image, which is combinedwith the information obtained by their intensityor colour --if neccessary. Two "post-processing" techniquesare proposed...

In this paper we propose a new method for image segmen- tation. The new algorithm is applied to t... more In this paper we propose a new method for image segmen- tation. The new algorithm is applied to the video segmenta- tion task, where the localization of moving objects is based on change detection. The change detection problem in the pixel domain is formulated by two zero mean Laplacian dis- tributions. The new method follows the concept of the well known Seeded Region Growing technique, while is adapted to the statistical description of change detection based seg- mentation, using Bayesian dissimilarity criteria in a way that leads to linear computational cost of growing. The segmentation algorithm is mainly based on change de- tection. The two classes of "changed"/"unchanged" pix- els are modeled by two Laplacian distributions. Let D = fd(s); s 2 Sg denote the gray level difference of each site s in the image grid S. The change detection problem consists of determining a binary label Θ(s) for each pixel s. We asso- ciate the random field Θ(s) with two possible ...

Eighth International Workshop on Image Analysis for Multimedia Interactive Services (WIAMIS '07), 2007

We propose a generic, unsupervised feature classification and image segmentation framework, where... more We propose a generic, unsupervised feature classification and image segmentation framework, where only the number of classes is assumed as known. Image segmentation is treated as an optimization problem. The framework involves block-based unsupervised clustering using k-means, followed by region growing in spatial domain. High confidence statistical criteria are used to compute a map of initial labelled pixels. A new region growing algorithm is introduced, which is named Independent Flooding Algorithm and computes a height per label for each one of the unlabeled pixels, using Bayesian dissimilarity criteria. Finally, a MRF model is used to incorporate the local pixel interactions of label heights and a graph cuts algorithm performs the final labelling by minimizing the underlying energy. Segmentation results using texture, intensity and color features are presented.

Signal Processing: Image Communication, 2001

This paper describes a semi-automatic method for moving object segmentation and tracking. This me... more This paper describes a semi-automatic method for moving object segmentation and tracking. This method is suitable when a few objects have to be tracked, while the camera moves and fixates on them. The user delineates approximately the initial locations in a selected frame and specifies the depth ordering of the objects to be tracked. First, motion-based segmentation is obtained through an initial application of a region growing algorithm. The partition map is sequentially tracked from frame to frame using motion compensation and location prediction. The segmentation map is obtained by the region growing algorithm. Translational motion is assumed for the moving objects, and local intensity or color average may be used as additional features. A post-processing procedure regularizes the object boundaries over time. r

EURASIP Journal on Advances in Signal Processing, 2003

EURASIP Journal on Advances in Signal Processing, 2002

IEEE Transactions on Image Processing, 2011

We propose a general purpose image segmentation framework, which involves feature extraction and ... more We propose a general purpose image segmentation framework, which involves feature extraction and classification in feature space, followed by flooding and merging in spatial domain. Region growing is based on the computed local measurements and distances from the distribution of features describing the different classes. Using the properties of the label dependent distances spatial coherence is ensured, since the image features are described globally. The distribution of the features for the different classes are obtained by blockwise unsupervised clustering based on the construction of the minimum spanning tree of the blocks' grid using the Mallows distance and the equipartition of the resulting tree. The final clustering is obtained by using the k-centroids algorithm. With high probability and under topological constraints, connected components of the maximum likelihood classification map are used to compute a map of initially labelled pixels. An efficient flooding algorithm is introduced, namely, Priority Multi-Class Flooding Algorithm (PMCFA), that assign pixels to labels using Bayesian dissimilarity criteria. A new region merging method, which incorporates boundary information, is introduced for obtaining the final segmentation map. Therefore, the merging stage is based on region features and edge localization. Segmentation results on the Berkeley benchmark data set demonstrate the effectiveness of the proposed methods.

ISPRS Journal of Photogrammetry and Remote Sensing, 2016

EURASIP Journal on Advances in Signal Processing, 2003

EURASIP Journal on Advances in Signal Processing, 2002

IEEE Transactions on Image Processing, 2011

Eighth International Workshop on Image Analysis for Multimedia Interactive Services (WIAMIS '07), 2007

Signal Processing: Image Communication, 2001

EURASIP Journal on Advances in Signal Processing, 2003

EURASIP Journal on Advances in Signal Processing, 2002

IEEE Transactions on Image Processing, 2011

We propose a general purpose image segmentation framework, which involves feature extraction and ... more We propose a general purpose image segmentation framework, which involves feature extraction and classification in feature space, followed by flooding and merging in spatial domain. Region growing is based on the computed local measurements and distances from the distribution of features describing the different classes. Using the properties of the label dependent distances spatial coherence is ensured, since the image features are described globally. The distribution of the features for the different classes are obtained by blockwise unsupervised clustering based on the construction of the minimum spanning tree of the blocks' grid using the Mallows distance and the equipartition of the resulting tree. The final clustering is obtained by using the k-centroids algorithm. With high probability and under topological constraints, connected components of the maximum likelihood classification map are used to compute a map of initially labelled pixels. An efficient flooding algorithm is introduced, namely, Priority Multi-Class Flooding Algorithm (PMCFA), that assign pixels to labels using Bayesian dissimilarity criteria. A new region merging method, which incorporates boundary information, is introduced for obtaining the final segmentation map. Therefore, the merging stage is based on region features and edge localization. Segmentation results on the Berkeley benchmark data set demonstrate the effectiveness of the proposed methods.