Mean Shift Clustering

Background: Understanding neural functions requires knowledge from analysing electrophysiological data. The process of assigning spikes of a multichannel signal into clusters, called spike sorting, is one of the important problems in such analysis. There have been various automated spike sorting techniques with both advantages and disadvantages regarding accuracy and computational costs. Therefore, developing spike sorting methods that are highly accurate and computationally inexpensive is always a challenge in the biomedical engineering practice.-New method: An automatic unsupervised spike sorting method is proposed in this paper. The method uses features extracted by the locality preserving projection (LPP) algorithm. These features afterwards serve as inputs for the landmark-based spectral clustering (LSC) method. Gap statistics (GS) is employed to evaluate the number of clusters before the LSC can be performed.-Results: The proposed LPP-LSC is a highly accurate and computationally inexpensive spike sorting approach. LPP spike features are very discriminative; thereby boost the performance of clustering methods. Furthermore, the LSC method exhibits its efficiency when integrated with the cluster evaluator GS.-Comparison with existing methods: The proposed method's accuracy is approximately 13% superior to that of the benchmark combination between wavelet transformation and superparamagnetic clustering (WT-SPC). Additionally, LPP-LSC computing time is six times less than that of the WT-SPC.-Conclusions: LPP-LSC obviously demonstrates a win-win spike sorting solution meeting both accuracy and computational cost criteria. LPP and LSC are linear algorithms that help reduce computational burden and thus their combination can be applied into real-time spike analysis.

This paper is mainly focussing on the area image database indexing. For this purpose a user interactive image segmentation algorithm is applied to the input image. After performing segmentation of the input image only the desired portion of the image i.e., the foreground is extracted. Interactive segmentation stage is followed by a feature extraction process that is performed for the image database as well as the region database. The features extracted include some colour features and texture features. Then using Manhattan distance a similarity measurement is done for images under same concept as well as images under different concept for both images in database as well as well as for segmented regions. It is found that user interactive image segmentation proves to be a very efficient method of indexing of image database.

This paper is mainly focusing on the area user interactive image segmentation. For this purpose a user interactive image segmentation algorithm is applied to the input image. After performing segmentation of the input image only the desired portion of the image i.e., the foreground is extracted. The segmentation method used in this project is semi-automated in nature. This algorithm begins with an initial over segmentation that is performed using mean shift clustering mechanism. Then the selected foreground and background regions are distinguished as white and black regions. Merging the unmarked regions as either foreground or background is done. Then a local neighbourhood classification is performed followed by the pruning step. The final result of this paper will isolate the foreground region as desired by the user from the background region. Foreground will be represented by white region and background by black region.

Robust vehicle tracking is essential in traffic
monitoring because it is the groundwork to higher level tasks
such as traffic control and event detection. This paper describes
a new technique for tracking vehicles with mean-shift using
a projective Kalman filter. The shortcomings of the meanshift
tracker, namely the selection of the bandwidth and
the initialization of the tracker, are addressed with a fine
estimation of the vehicle scale and kinematic model. Indeed, the
projective Kalman filter integrates the non-linear projection of
the vehicle trajectory in its observation function resulting in an
accurate localization of the vehicle in the image. The proposed
technique is compared to the standard Extended Kalman filter
implementation on traffic video sequences. Results show that
the performance of the standard technique decreases with the
number of frames per second whilst the performance of the
projective Kalman filter remains constant.

Image segmentation is a critical step in computer vision tasks constituting an essential issue for pattern recognition and visual interpretation. In this paper, we study the behavior of entropy in digital images through an iterative algorithm of mean shift filtering. The order of a digital image in gray levels is defined. The behavior of Shannon entropy is analyzed and then compared, taking into account the number of iterations of our algorithm, with the maximum entropy that could be achieved under the same order. The
use of equivalence classes it induced, which allow us to interpret entropy as a hyper-surface in real mdimensional space. The difference of the maximum entropy of order n and the entropy of the image is used to group the the iterations, in order to caractrizes the performance of the algorithm.

Abnormal growth of tissue in the brain which results in tumor in the brain called brain tumor which causes death of human being. Hence the early detection of tumor is necessary to survive a human being. An automatic pc oriented diagnosing method has designed in order to enhance accuracy of detection and classification of tumor in a brain. Here we are using PNN classifier along with mean shift clustering method for detection and classification of different stages of brain tumor such as Normal, Benign stage and malignant stage. PNN (Probabilistic Neural Network) is a learning machine is basically used with an extension of various image classifications based on Training and Testing networks. Probabilistic Neural Network results in accurate & fast classifying tool for detection and classification of tumor.