Combining Classifiers: from the creation of ensembles - ICMC
... • training data manipulation: training an ensemble of
classifiers using different training sets by splitting ,
by using cross-validation ensembles (used with small
subsets) , bagging  and boosting .
• input features manipulation: training an ensemble of
classifiers using different su ...
Ancient Tamil Script Recognition from Stone Inscriptions Using Slant
... characters whilst the white pixels are used to represent
A. Core-region Detection
Our approach is based on enhanced horizontal density
histogram embedded with quantile features to detect core
region of stone inscription images. Quantiles are points taken
at regular intervals from the cum ...
Robust Continuous Collision Detection Between
... spaces in between stacked objects. On the other hand, late detection of collisions may cause
objects to interpenetrate or overlap which causes rendering artifacts such as Z-fighting where
rendered pixels on screen rapidly change color as the view point changes gradually. Not only
accuracy, but also ...
Computational Intelligence in Intrusion Detection System
... Intrusion detection system (IDS) is a major research problem in network
security, its goal is to dynamically identify unusual access or attacks to secure
the networks. Network intrusion detection systems (NIDS) is a valuable
tool for the defense in depth of computer networks. It looks for known
or p ...
Following non-stationary distributions by controlling the
... sudden change. So the challenge is to exploit the properties of algorithms like
GNG, while controlling the additions and removals of neurons, in order to
stick to the changes in the distribution. This requires computing some kind
of statistics for each prototype, in order to detect when it needs to ...
Fast Visualization of Object Contours by Non
... high intensity pixel there are several high opacity samples
along the corresponding viewing ray resulting in high accumulated opacity. Therefore, the fine details are less visible.
This version is appropriate for data sets which contain many
high frequency regions. In this case only the main charact ...
... identification and recognition system. The system consists of three main parts; are image acquisition,
image processing and "identification". The fingerprint images captured sensor format and stored in the
database in the process of image acquisition. This image is the commitment to share 500 DPI. T ...
On Constrained Optimization Approach to Object
... from the interior, as applied to image segmentation,
would provide an invaluable insight in
distinguishing the connected segments. For
instance, if many seeds grow into many plants, any
leaf in dense foliage can still be traced to its own
stem, even when the connection to its origin may be
Nonparametric Curve Extraction Based on Ant Colony System Qing Tan Qing He
... sensing images and other crack detection, curve equations
are usually not known a priori before recognition. HT
methods do not work in these situations.
Extracting curves from a binary image means to select a
subset of pixels (real target pixels) from all nonzero pixels.
So every pixel sequence can ...
Predicting Classifier Combinations
... lead to the following issues: If multiple combinations achieve the highest accuracy, the label includes
only one of them and predicting any other combination with the same accuracy will lead to an error.
Also, predicting a sub-optimal combination with only
a slightly decreased accuracy as compared t ...
Learning Visual Representations for Perception
... We evaluated our system on an abstract task that closely parallels a
real-world, reactive navigation scenario (Fig. 3). The goal of the agent
is to reach one of the two exits of the maze as fast as possible. The
set of possible locations is continuous. At each location, the agent
has four possible a ...
Analysis of Machine Learning Techniques for Intrusion Detection
... Intrusions. There are many machine learning techniques used
in Intrusion Detection System and they comprised single,
hybrid and ensemble classifiers. Many resources have been
used on various machine learning techniques. These techniques
work very well for IDS but it is known that there is not even a ...
An Analytical Study of the Remote Sensing Image Classification
... Firefly Algorithm etc. Swarm Intelligence is a global research area to improve the optimization of various soft computing and
nature inspired techniques. In this paper, we are using these swarm intelligence based techniques for the accurately
classification of land cover features. For this classific ...
A Stochastic Algorithm for Feature Selection in Pattern Recognition
... discriminant properties . In a recent work of Fleuret (2004), the author suggests to use mutual
information to recursively select features and obtain performance as good as that obtained with a
boosting algorithm (Friedman et al., 2000) with fewer variables. Weston et al. (2000) and Chapelle
et al. ...
An Information Theoretic Approach to Reflectional Symmetry Detection
... Seldom does symmetry occur by accident. If two regions
are symmetric it is highly probable that these regions are
related in the real world, or even that they belong to the same
object. Therefore, by detecting symmetry it is possible to start
grouping or segmenting the image without prior knowledge
IOSR Journal of Electronics and Communication Engineering (IOSR-JECE)
... image was not used then the features of the same leaf vary with different sizes images. Since, it was the same
leaf, the value of all the features were expected to be same. The values for these features vary for different
scaled versions of the same leaf image. In this research, features used were s ...
Local Scale Control for Edge Detection and Blur Estimation
... decreases with time (smoothing), above which the gradient
increases with time (edge enhancement). Unfortunately,
there is no principled way of choosing this threshold even
for a single image, since important edges generate a broad
range of gradients, determined by contrast and degree of
blur. Sensor ...
Preparation of Papers in Two-Column Format for
... algorithm. A two layered feed-forward neural network created is then used as a classifying tool for the input images. The
consistency of the results is demonstrated by the median value. The performance achieved here is 82%. The method is not
able to treat situations when the eyes are closed. Strong ...
Intrusion detection using clustering
... S. Otherwise the cluster which gives minimum distance
between the data point and the cluster is found, next the
distance is compared the result with the cluster radius
threshold, if it is less than the defined threshold then the data
point is added to the cluster otherwise new cluster is created
Using Natural Image Priors
... Figure 1.2: Left: The energy function defined over derivative outputs that was learned
by Zhu and Mumford . Right: The potential function e−E(x) . The blue curves
show the original function and the red curves are approximations as a mixture of 50
Gaussians. In this work we use this mixture appro ...
Survey on Remotely Sensed Image Classification
... problem effectively when converting the qualitative case to
quantitative ones. The example study shows that the SVMFAHP model is feasible and effective. The research can
provide decision-making for enterprises to select 3PL
Giorgos Mountrakis, Jungho Im, Caesar Ogole 
studied a wide ra ...
Intrusion Detection using Fuzzy Clustering and Artificial Neural
... This paper presents the outline of a hybrid
Artificial Neural Network (ANN) based on fuzzy
clustering and neural networks for an Intrusion
Detection System (IDS). While neural networks
are effective in capturing the non-linearity in data
provided, it also has certain limitations including
the requir ...
Painting-to-3D Model Alignment Via Discriminative Visual
... matching was also used for aligning paintings to 3D meshes
reconstructed from photographs [Russell et al. 2011]. However,
contours extracted from paintings and real-world 3D meshes
obtained from photographs are noisy. As a result, the method
requires a good initialization with a close-by viewpoint. ...
Distractor Quality Analyze In Multiple Choice Questions
language" and "byte code" to the document image of article A,
considering them synonymous with the keyword "java". The
same rule applies to other keywords: If they have synonyms,
then all they need to be added in the document image. But
there is another way. Its essence i ...
Histogram of oriented gradients
The histogram of oriented gradients (HOG) is a feature descriptor used in computer vision and image processing for the purpose of object detection. The technique counts occurrences of gradient orientation in localized portions of an image. This method is similar to that of edge orientation histograms, scale-invariant feature transform descriptors, and shape contexts, but differs in that it is computed on a dense grid of uniformly spaced cells and uses overlapping local contrast normalization for improved accuracy.Navneet Dalal and Bill Triggs, researchers for the French National Institute for Research in Computer Science and Automation (INRIA), first described HOG descriptors at the 2005 Conference on Computer Vision and Pattern Recognition (CVPR). In this work they focused on pedestrian detection in static images, although since then they expanded their tests to include human detection in videos, as well as to a variety of common animals and vehicles in static imagery.