Konstantia Moirogiorgou

Feature analysis on river flow video data for floating tracers detection

This study focuses on a significant water quality problem that can be posed as the calculation of the distribution of suspended sediments in rivers. More specifically, we propose a method that performs tracking and motion estimation on river floating sediment tracers. The current work analyzes a river flow video sequence and isolates the sediment information in order to determine the temporal extend of the suspended sediment distribution. The method is based on a combination of image processing techniques and is performed through hue and intensity analysis. The results present the river sediment tracers isolated from the river water flow and the river background as well as the river floating tracers velocity vector field.

Extracting reliable gene expression signatures through Stable Bootstrap Validation.

Identification of candidate genes responsible for specific phenotypes, such as cancer, has been a major challenge in the field of bioinformatics. Given a DNA Microarray dataset, traditional feature selection methods produce lists of candidate genes which vary significantly under variations of the training data. That instability hinders the validity of research findings and raises doubts about the reliability of such methods. In this study, we propose a framework for the extraction of stable genomic signatures. The proposed methodology enforces stability at the validation step, independent of the feature selection and classification methods used. The statistical significance of the selected gene set is also assessed. The results of this study demonstrate the importance of stability issues in genomic signatures, beyond their prediction capabilities.

River flow estimation using video data

An image-based framework for river flow monitoring based on a statistical estimation technique for fluid flow estimation is presented. This approach uses subsequent gray-scale video frames along with a statistical estimation method to extract the optical flow. An average velocity estimate is computed using the velocity vectors of the main motion trend, which is extracted using classification methods. The corresponding real-world surface velocity is computed using velocity-area transformations. The use of only two subsequent video frames and the lack of tracers in the flow are the key features of this technique in order to extract an accurate estimate of the real surface velocity. We compare our real-world surface velocity estimate with traditional current meter measurements, made on the site of Pinios river, Thessaly, Greece using the Q-liner 2 Doppler device.

Stereo System for Remote Monitoring of River Flows

In this article we present a video-based method for river flow monitoring. The proposed method aims at deriving efficient approximations of the river velocity using natural formations on the river surface. In order to overcome peculiarities of the flow, we propose to uniformly exploit all such structures that appear locally with short temporal duration. Towards this direction we explore the expanded capabilities of a stereoscopic camera layout with the dual observation fields and the potential of reverting projective deformations. By mapping to world coordinates, all spatial locations in the video reflect velocity as a uniform field, except for local flow variations. The velocity estimation is performed by computing the optical flow using a series of video frames, combining the information of the views of both cameras. The novelty of the proposed river flow estimation scheme lies on the fact that the accuracy of motion estimation is increased due to the use of the complementary views, which also enables the transition from a 2-Dimensional image-based velocity estimate to 3-Dimensional estimates. The estimated optical velocity is back-projected to the real world coordinates using the parameters extracted using the stereoscopic layout. The results on simulated and real conditions demonstrate that the proposed method is efficient in the estimation of the surface velocity and robust against locally disappearing formations, since it can compensate for a loss with other formations active in the field of view.

3D Building Reconstruction using Stereo Camera and Edge Detection

Three dimensional geo-referenced data for buildings are very important for many applications like cadastre, urban and regional planning, environmental issues, archaeology, architecture, tourism and energy. The acquisition and update of existing databases is time consuming and involves specialized equipment and heavy post processing of the raw data. In this study we propose a system for urban area data based on stereo cameras for the reconstruction of the 3D space and subsequent matching with limited geodetic measurements. The proposed stereo system along with image processing algorithms for edge detection and characteristic point matching in the two cameras allows for the reconstruction of the 3D scene in camera coordinates. The matching with the available geodetic data allows for the mapping of the entire scene on the word coordinates and the reconstruction of real world distance and angle measurements.

Automated Waterline Extraction for Optimal Land Use - A Case Study in Crete

Coastal areas in Greece play an important role to the economic growth of the country. Especially in the islands as a major tourist destination, new tourist infrastructure is continuously being built. To optimize land use and protect current infrastructure, we need to properly define the waterline. In addition, the waterline is changing dynamically due to climate change and weather conditions, so it has to be monitored. The proposed framework is focused on the automatic extraction of waterline from aerial images using advanced image processing techniques. We plan to integrate the Greek cadastral data in the framework and along with spatial data analysis to be able to proceed with both coastal and cadastral data. Additional functionalities are provided to the planners, through metrics for distance and surface area calculations, in order to extract useful information for sustainable region planning of the coastal area. The results can be visualized in the images with the metrics

Color characteristics for the evaluation of suspended sediments

This study focuses on a significant issue of the environmental monitoring application area, which is the suspended sediment concentration estimation. More specifically, the purpose of the current work is to provide a new non-intrusive way to estimate the suspended sediment (SS) distribution. The proposed methodology uses the color characteristics of river flow images and provides a high correlation factor with the suspended sediment measurements. In our opinion, the importance of the current work derives from the fact that it provides an alternative and effective way of estimating SS distribution rather as opposed to the conventional method that requires human presence, especially if we consider the difficulty of taking measurements of the river pollution during flush flood events when the sediment distribution is increased and is directly related to water quality.

Lung tissue evaluation detecting and measuring morphological characteristics of cell regions

The goal of this study is to develop an automated, accurate and time efficient image processing algorithmic scheme, capable of segmenting lung tissue slides and quantitatively detecting any possible morphological characteristic that may differentiate healthy cells from adenocarcinoma. Microscopy images are segmented into the key regions via a proposed clever, sequential fusion methodology, combining image clustering, the watershed transform and mathematical morphology and analyzed utilizing an innovative tissue evaluation approach based on quantitative assessments of the extracted cell regions shape and size. The preliminary results of this work indicate that it is possible to discriminate healthy cells from cancerous ones considering their overall morphology within the tissue and measuring possible indices that may reveal an evolving neoplasia, a tumor growth or a malfunction in cell proliferation. Applying the proposed method to a much larger and more variform dataset is our next plan for the upcoming future in order to validate and ensure the robustness and accuracy of the proposed classification scheme, making it an extremely valuable assisting tool for medical experts for cancer diagnosis and prognosis.

Statistical Estimation of Fluid Flow: An Image Restoration Approach

This paper focuses on Fluid Motion-Field Estimation from video data, which is a useful but challenging problem in environmental monitoring. Rivers are often monitored by flashy hydrographs that exhibit characteristic response times ranging from minutes to hours. In order to estimate the river discharge during a flush flood event, the temporary motion vector field of the river surface is needed. This paper presents a new approach in statistical estimation of fluid flow that calculates a local flow probability distribution function in the frequency domain. Our work improves upon the inefficiencies of spatial estimation of the auto-regressive STAR model and converts motion estimation into a restoration problem, where the local field can be computed fast in the frequency domain, while various natural constraints can be taken into account within the inversion strategy of the motion estimation process.

A Real-Time System for HDTV Video Transmission over IP Networks

We present an end-to-end server–client system for high-resolution video transmission over IP networks. The system supports communication of compressed information, user operations through a fully functional user interface as well as detection and tracking of moving objects/people. It relies on a communication subsystem which has been developed for real time transmission of MJPEG / MJPEG2000 compressed video information over erroneous out-of-order packet-based local and wide area networks. The images are at the HDTV standard resolution. The server grabs frames from a camera-link sensor, encodes the raw video, places the encoded bit stream into packets and transmits them over network. The client receives and reorders the packets and then displays the video data. Each user can request different focus windows from the server according to the viewpoint of interest. Each encoded frame is filled with restart markers in order to face up the case of a frame been lost.