Pooja Agrawal

Segmentation of Handwritten Hindi Text: A Structural Approach

This paper makes an attempt to segment the handwritten Hindi words. The problem of segmentation is compounded by the possible presence of modifiers known as matras on all sides of the basic characters and due to the uncertainty introduced in the character shapes by way of different writing styles. We have devised a structural approach to capture the similarities and differences between structure classes. The segmentation is performed in hierarchical order: 1) Separating the upper modifiers and header line from the character, 2) Detecting and then segmenting lower modifiers from the characters, 3) Identifying whether a character is conjunct or not, 4) Categorization of top modifiers based on Check_point, Mid_point and Touching_points. The segmentation accuracy has been found to be around 78%. Some general conditions are applied for separating modifiers from the characters. But certain words cannot be segmented because they violate the general conditions. However, specifics are not dealt with in this paper because such an attempt requires an exhaustive study on a large database that is not available presently.

MRI brain image segmentation for spotting tumors using improved mountain clustering approach

This paper presents improved mountain clustering technique based MRI (magnetic resonance imaging) brain image segmentation for spotting tumors. The proposed technique is compared with some existing techniques such as K-Means and FCM, clustering. The performance of all these clustering techniques is compared in terms of cluster entropy as a measure of information and also is visually compared for image segmentation of various brain tumor MRI images. The cluster entropy is heuristically determined, but is found to be effective in forming correct clusters as verified by visual assessment.

Vision Based Obstacle Detection and Avoidance for UAVs Using Image Segmentation

The present work proposes a vision based guidance scheme for an unmanned aerial vehicle (UAV) navigating through urban environments. Optical flow of image feature is considered to segment obstacles from the image. With the segmented obstacles, a collision cone based guidance law is proposed for avoidance. Detecting open space between segmented obstacles, a passage following guidance law is also presented for intelligent decision making. Simulations are carried out in a 3D environments created in VRML toolbox of MATLAB R ©. Results shows a much improved performance as compared to existing optical flow based obstacle avoidance methods.

Medical Image Segmentation Using Improved Mountain Clustering Approach

This paper presents Improved Mountain Clustering (IMC) based medical image segmentation. Proposed technique is a more powerful approach for X-Ray image based diagnosing diseases like lung cancer and tuberculosis. The IMC based segmentation approach was applied on lung X-Ray images and compared with some existing techniques such as K-Means and FCM based segmentation approaches. The performance of all these segmentation approaches is compared in terms of cluster entropy as a measure of information. The segments obtained from the methods have been verified visually.

Adaptive backstepping sliding mode control based on nonlinear disturbance observer for trajectory tracking of robotic manipulator

This paper proposes and analyzes two different control techniques to remove the effect of lumped uncertainties and disturbances for trajectory tracking by robotic manipulator. These techniques are a) Adaptive backstepping sliding mode control and b) Nonlinear disturbance observer based backstepping sliding mode control. Adaptive backstepping sliding mode control estimates the system uncertainties and disturbance using an adaptive law. Lyapunov theory is used to define the adaptive law for the convergence of tracking error. The second technique initially estimates the unknown external disturbances using non-linear disturbance observer and then generates control input using beckstepping sliding mode controller. Backstepping sliding mode ensures the sliding surface to be chattering free and improves convergence rate in finite-time. The stability of system is analyzed using Lyapunov theory for both the techniques. Simulation results show the effectiveness and robustness of proposed techniques FOR trajectory tracking.

Color Segmentation Using Improved Mountain Clustering Technique Version-2

This paper proposes a heuristically optimized version of Improved Mountain Clustering (IMC) Technique referred to as IMC-2. IMC-2 provides better quality clusters measured in terms of Global Silhouette and Separation indices as measures of information. The IMC-2 based color segmentation approach has been applied to various categories of images including face, stripes and grayscale images and compared with some extensively used clustering techniques such as K-means and FCM. The color segmentation performance has been compared on widely used and accepted validation indices, Global Silhouette Index and Separation Index. The color segments or clusters obtained have been verified visually and validated quantitatively.

Fuzzy rule based unsupervised approach for salient gene extraction

This paper presents a novel fuzzy rule based gene ranking algorithm for extracting salient genes from a large set of microarray data which helps us to reduce computational efforts towards model building process. The proposed algorithm is an unsupervised approach and does not require class information for gene ranking and Microarray data has been used to form a set of robust fuzzy rule base which helps us to find salient genes based on its average relevance with already formed fuzzy rules in rule base. Fuzzy rule based ranking has been carried out to select salient genes based on their average firing strength in order of high relevancy and only top ranked genes are utilized to classify normal and cancerous tissues for a carcinoma dataset [1]. Result validate the effectiveness of our gene ranking method as for the same no. of genes, our ranking scheme helps to improve the classifier performance by selecting better salient genes.

A Composite Guidance Strategy for Optical Flow Based UAV Navigation

Abstract This work presents an optical flow composite guidance strategy for UAV navigation in unknown outdoor environments while seeking a predefined goal point. The proposed guidance logic is based on the relative optical flow in the right, left, and center regions of image. The resulting guidance strategy command uses balance and inverse strategies to navigate through different obstacle scenarios. Simulation studies are carried out in 3 D environment created in Virtual Reality Modelling Languages toolbox, MATLAB ® . In terms of initial heading and control efforts, the proposed composed strategy shows a better performance as compared to individual balance and inverse strategies.

Sonography as an objective tool for monitoring serial corrections and detecting spurious corrections in clubfoot: A review

Ultrasonography is an emerging tool for monitoring clubfoot correction and for early diagnosis of spurious correction and of deformity recurrence. Sonography is widely available, inexpensive and has dynamic capability and can visualize tarsals in infants accurately.

Fuzzy rule based unsupervised approach for gene saliency.

Clinical background This abstract presents a novel fuzzy rule based gene ranking algorithm for extracting salient genes from a large set of microarray data which helps us to reduce computational efforts towards model building process. The proposed algorithm is an unsupervised approach and does not require any prior class information for gene ranking and microarray data has been used to form a set of robust fuzzy rule base which helps us to find salient genes based on its average relevance with already formed fuzzy rules in rule base. Fuzzy rule based ranking has been carried out to select salient genes based on their average firing strength (i.e. average true value after all the fuzzy rules applied) in order of high relevancy and only top ranked genes are utilized to classify normal and cancerous tissues for a carcinoma dataset [1]. Results validate the effectiveness of our gene ranking method as for the same no. of genes, our ranking scheme helps to improve the classifier performance by selecting better salient genes. In our case study the performance comparison for five top ranked genes is given in Table 1. Conclusion Results of classifiers in terms of correct rate (Table 1) show that the proposed fuzzy rule based gene ranking scheme outperforms t-test based ranking schemes.