Detection Rate

Aim  Automatic CT dataset classification is important to efficiently create reliable database annotations, especially when large collections of scans must be analyzed. Method  An automated segmentation and labeling algorithm was developed based on a fast patient segmentation and extraction of statistical density class features from the CT data. The method also delivers classifications of image noise level and patient size. The approach

Network intrusion detection often finds a difficulty in creating classifiers that could handle unequal distributed attack categories. Generally, attacks such as Remote to Local (R2L) and User to Root (U2R) attacks are very rare attacks and even in KDD dataset, these attacks are only 2% of overall datasets. So, these result in model not able to efficiently learn the characteristics of rare categories and this will result in poor detection rates of rare attack categories like R2L and U2R attacks. We even compared the accuracy of KDD and NSL-KDD datasets using different classifiers in WEKA.

Acute kidney injury (AKI) is associated with significantly increased morbidity and mortality. To provide a uniformly accepted definition, the RIFLE classification was introduced by the Acute Dialysis Quality Initiative, recently modified by the Acute Kidney Injury Network (AKIN), suggesting staging of AKI based on dynamic changes within 48 h. This study compares these two classification systems with regard to outcome. Cohort analysis of SAPS 3 database. Sixteen thousand seven hundred and eighty-four ICU patients from 303 ICUs were analysed. Classification was performed according to RIFLE (Risk, Injury, Failure) or according to AKIN (stage 1, 2, 3) without including a requirement of renal replacement therapy in the analysis. Changes of serum creatinine as well as urinary output were assessed for both AKIN and RIFLE during the first 48 h of ICU admission. Primary endpoint was hospital mortality. Incidence of AKI in our population of critically ill patients was found to range between 28.5 and 35.5% when applying AKIN and RIFLE criteria, respectively, associated with increased hospital mortality averaging 36.4%. Observed-to-expected mortality ratios revealed excess mortality conferred by any degree of AKI increasing from 0.81 for patients classified as non-AKI up to 1.31 and 1.23 with AKIN stage 3 or RIFLE Failure, respectively. AKIN misclassified 1,504 patients as non-AKI compared to RIFLE which misclassified 504 patients. Acute kidney injury classified by either RIFLE or AKIN is associated with increased hospital mortality. Despite presumed increased sensitivity by the AKIN classification, RIFLE shows better robustness and a higher detection rate of AKI during the first 48 h of ICU admission.

Network intrusion detection often finds a difficulty in creating classifiers that could handle unequal distributed attack categories. Generally, attacks such as Remote to Local (R2L) and User to Root (U2R) attacks are very rare attacks and even in KDD dataset, these attacks are only 2% of overall datasets. So, these result in model not able to efficiently learn the characteristics of rare categories and this will result in poor detection rates of rare attack categories like R2L and U2R attacks. We even compared the accuracy of KDD and NSL-KDD datasets using different classifiers in WEKA.

In this paper QRS complex detection algorithms based on the first and second derivatives have been studied and implemented. The threshold values for detecting R-peak candidate points mentioned in previous work have been modified for accuracy point of view. The derivative based QRS detection algorithms have been found not only computationally simple but exceptionally effective also on variety of ECG database that includes highly noisy and arrhythmic ECG signals. This is indicated by an average detection rate of over 98% obtained through the modified threshold values even for the challenging ECG test sets.

Automation of oral reading assessment and of feedback in a read-ing tutor is a very challenging task. This paper describes our re-search aiming at developing such automated systems. First topic is the recording and annotation of CHOREC, the Flemish database of children's oral ...

A method for blink detection from video sequences gathered with a commercial camera is presented. This is used as a view-based remote eye gaze tracker (REGT) component performing two relevant functions, i.e. initialization and automatic updating in case of tracking failures. The method is based on frame differencing and eyes anthropometric properties. It has been tested on a publicly available database and results have been compared with algorithms found in literature. The obtained average true prediction rate is higher than 95%. The robustness of the automatic tracking failure detection has been tested on a set of experimental trials in different conditions, and yielded detection rates around 98%. The computational cost of the processing allows the blink detection algorithm to work in real time at 30 fps. The obtained results are in favour of combining blink detection with gaze mapping for the development of a robust view-based remote eye-gaze tracker to be introduced in different HCI contexts, specifically in the assistive technology framework.

Illegal logging is a major environmental and economic problem, and exceeds in some countries the amounts of legally harvested timber. In Eastern Europe and the former Soviet Union, illegal logging increased and reforestation on abandoned farmland was widespread after the breakdown of socialism, and the region's forest cover trends remain overall largely unclear. Our goal here was to map forest cover change and to assess the extent of illegal logging and reforestation in the Ukrainian Carpathians. We used Landsat TM/ETM+ images and Support Vector Machines (SVM) to derive forest change trajectories between 1988 and 2007 for the entire Ukrainian Carpathians. We calculated logging and reforestation rates, and compared Landsat-based forest trends to official statistics and inventory maps. Our classification resulted in reliable forest/non-forest maps (overall accuracies between 97.1%–98.01%) and high clear cut detection rates (on average 89.4%). Forest cover change was widespread in the Ukrainian Carpathians between 1988 and 2007. We found forest cover increase in peripheral areas, forest loss in the interior Carpathians, and increased logging in remote areas. Overall, our results suggest that unsustainable forest use from socialist times likely persisted in the post-socialist period, resulting in a continued loss of older forests and forest fragmentation. Landsat-based forest trends differed substantially from official forest resource statistics. Illegal logging appears to have been at least as extensive as documented logging during the early 1990s and so-called sanitary clear-cuts represent a major loophole for overharvesting and logging in restricted areas. Reforestation and illegal logging are frequently not accounted for in forest resource statistics, highlighting limitations of these data. Combating illegal logging and transitioning towards sustainable forestry requires better monitoring and up-to-date accounting of forest resources, in the Carpathians and elsewhere in Eastern Europe, and remote sensing can be a key technology to achieve these goals.