Nurcan Durak

Coronal loop detection from solar images

In this paper, we make an overview of a methodology for the automatic retrieval of images with coronal loops from the solar image data captured by the extreme-ultraviolet imaging telescope (EIT) onboard the spacecraft SOHO (Solar and Heliospheric Observatory). Our image retrieval system provides relevant data to astrophysicists who need such data to study the coronal heating problem. As part of building this system, we investigated various image preprocessing techniques, image based features, and classifiers to automatically detect coronal loops and to indicate their locations on the images. Despite many challenges related to the coronal loop characteristic, we obtained promising results, namely, 78% precision and 80% recall in loop retrieval.

Feature Exploration for Mining Coronal Loops from Solar Images

Coronal loops are especially important in analyzing some important phenomena related to the Sun such as the controversial coronal heating problem. The analysis requires astrophysicists to manually sift through thousands of images in order to acquire images containing coronal loops. Thus, the motivation to detect these loops automatically. Since coronal loops do not have a perfect shape and are easy to confuse with other solar events, feature selection to learn characteristics of loops requires special care. In this study, we explore standard image features as well as specialized image features considering coronal loop characteristics. Our experiments confirm the success of our explored features in coronal loop detection.

Extracting salient contour groups from cluttered solar images via Markov Random Fields

We propose a method to extract salient contour groups from cluttered regions using Markov Random Fields. Our technique delineates smooth, long, and elliptical curves out of clutter. To extract salient curves, we use the following perceptual rules: smoothness, proximity, co-circularity, co-elliptic, and length. Our method consists of the following steps: obtaining individual smooth contours along their saliency measures; starting from the most salient contour search for possible grouping options for each contour; continuing the grouping until an optimum solution is reached. We introduce circularity and saliency measures for open curves. We applied our method to discern coronal loops from cluttered solar images, and were able to successfully obtain the entire set of desired coronal loops while eliminating any cluttered background.

Principal Contour Extraction and Contour Classification to Detect Coronal Loops from the Solar Images

In this paper, we describe a system that determines coronal loop existence from a given Solar image region in two stages: 1) extracting principal contours from the solar image regions, 2) deciding whether the extracted contours are in a loop shape. In the first stage, we propose a principal contour extraction method that achieves 88% accuracy in extracting the desired contours from the cluttered regions. In the second stage, we analyze the extracted contours in terms of their geometric features such as linearity, elliptical features, curvature, proximity, smoothness, and corner points. To distinguish loop contours from the other forms, we train an Adaboost classifier based C4.5 decision tree by using geometric features of 150 loop contours and 250 non-loop contours. Our system achieves 85% F1-Score from 10-fold cross validation experiments.

Multimodal video database modeling, querying and browsing

In this paper, a multimodal video indexing and retrieval system, MMVIRS, is presented. MMVIRS models the auditory, visual, and textual sources of video collections from a semantic perspective. Besides multimodality, our model is constituted on semantic hierarchies that enable us to access the video from different semantic levels. MMVIRS has been implemented with data annotation, querying and browsing parts. In the annotation part, metadata information and video semantics are extracted in hierarchical ways. In the querying part, semantic queries, spatial queries, regional queries, spatio-temporal queries, and temporal queries have been processed over video collections using the proposed model. In the browsing parts, video collections are navigated using category information, visual and auditory hierarchies.

Mining Coronal Loops in Solar Images from the SOHO collection

We present our preliminary findings as part of a new data mining application aiming at the automatic detection of images with coronal loops from one of NASAs solar image databases, known as EIT. Coronal loops are immense arches of hot gas on the surface of the Sun, thought to be jets of hot plasma flowing along in the alleys between the strong coronal magnetic fields. We use various data mining techniques including combining crisp and fuzzy classifiers for automated detection of blocks extracted from EIT solar images. Our data mining and retrieval system helps provide relevant data to astrophysicists who need such data to study the solar corona, and whose work is traditionally hindered by the need to manually sift through thousands of images in order to locate the very few that are useful for further analysis. Our data-driven approach is distinct from related image processing based approaches that cannot scale to large image databases because they rely mostly on semi-automated detection and on heavy and computationally intensive local shape analysis.