Showing papers by "Anupam Agrawal published in 2015"

Vision based hand gesture recognition for human computer interaction: a survey

Abstract: As computers become more pervasive in society, facilitating natural human---computer interaction (HCI) will have a positive impact on their use. Hence, there has been growing interest in the development of new approaches and technologies for bridging the human---computer barrier. The ultimate aim is to bring HCI to a regime where interactions with computers will be as natural as an interaction between humans, and to this end, incorporating gestures in HCI is an important research area. Gestures have long been considered as an interaction technique that can potentially deliver more natural, creative and intuitive methods for communicating with our computers. This paper provides an analysis of comparative surveys done in this area. The use of hand gestures as a natural interface serves as a motivating force for research in gesture taxonomies, its representations and recognition techniques, software platforms and frameworks which is discussed briefly in this paper. It focuses on the three main phases of hand gesture recognition i.e. detection, tracking and recognition. Different application which employs hand gestures for efficient interaction has been discussed under core and advanced application domains. This paper also provides an analysis of existing literature related to gesture recognition systems for human computer interaction by categorizing it under different key parameters. It further discusses the advances that are needed to further improvise the present hand gesture recognition systems for future perspective that can be widely used for efficient human computer interaction. The main goal of this survey is to provide researchers in the field of gesture based HCI with a summary of progress achieved to date and to help identify areas where further research is needed.

Target detection in SAR images using SIFT

Abstract: Target detection in synthetic aperture radar (SAR) images which are affected by speckle noise is a challenging task. An algorithm for automatic target detection in SAR images is proposed in this research work. In the first step, moving and stationary target acquisition and recognition (MSTAR) images are segmented and passed through multiple preprocessing stages (histogram equalization, dilation, position normalization). In the next step, feature extraction based on SIFT is performed. The extracted features from testing images are matched with the features extracted from training images. Thus, the classification of the targets is performed. The results obtained and the comparison with existing algorithms, both are sufficient enough to prove that the proposed algorithm is robust and effective.

Real Time Hand Detection & Tracking for Dynamic Gesture Recognition

Abstract: In the recent year gesture recognition has become the most intuitive and effective communication technique for human interaction with machines. In this paper we are going to work on hand gesture recognition and interpret the meaning of it from video sequences. Our work takes place in following three phases: 1. Hand Detection & Tracking 2. Feature extraction 3. Gesture recognition. We have started proposed work with first step as applying hand tracking and hand detection algorithm to track hand motion and to extract position of the hand. Trajectory based features are being drawn out from hand and used for recognition process and hidden markov model is being design for each gesture for gesture recognition. Hidden Markov Model is basically a powerful statistical tool to model generative sequences. Our method is being tested on our own data set of 16 gestures and the average recognition rate we have got is 91%. With proposed methodology gives the better recognition results compare with the traditional approaches such as PCA, ANN, SVM, DTW and many more.

Posture recognition for safe driving

Abstract: Recent evolution in Human-Computer Interaction automation grant permission for the integration between the driver's nature and its causes of accidents. To make an automated driver assistance system, the postures of the driving person have to be analysed and recognized efficiently. In this proposed work an automated driving assistance system is developed based on six different postures i.e Operating the steering wheel, Maintaining the shift lever, Eating during driving, Talking over the mobile, Controlling the Dash Board and Partial sleeping while driving (Drowsiness). In the mentioned six postures, three are considered as safe activities and the other three as unsafe activities. The proposed method starts with the feature extraction from sequence of input images followed by the classification process to recognize the activities. To ensure the effectiveness of the proposed system, an inhouse data set has been created and utilized for analysis. From the experimental results it has been observed that proposed system outperforms to the existing systems in terms of recognition accuracy which is 92.9 %.

Palatal Patterns Based RGB Technique for Personal Identification

Abstract: Biometric system is an alternative way to the traditional identity verification methods. This research article provides an overview of recently / currently used single and multiple biometrics based personal identification systems which are based on human physiological (such as fingerprint, hand geometry, head recognition, iris, retina, face recognition, DNA recognition, palm prints, heartbeat, finger veins, footprints and palates) and behavioral (such as body language, facial expression, signature verification and speech recognition) characteristics. This paper focuses on RGB based palatal pattern analysis of persons and the proposed technique uses RGB values with silhouette computes of palatal patterns for identifying a person. We have tested our proposed technique for palatal patterns of 50 persons including males & females and it is observed that RGB values based silhouette technique are accurately identifying the persons on the basis of their palatal patterns. For each person seven palatal images were taken. Out of these seven palatal images, four images were used for training dataset and last three palatal patterns were used for identifying the persons. The proposed technique is reliable & secure and it is a foolproof method which is clearly differentiating the persons on the basis of their palatal patterns.

Evaluation Schema for SAR image Segmentation based on Swarm Optimization in Neutrosophic Domain

Abstract: om Abstract - The present paper proposes an evaluation schema for SAR image Segmentation. The segmentation of SAR images becomes a crucial step because the segmented results will be used as the input for the post- processing like target detection, change detection applications. The preprocessing task segmentation is highly influenced by inherent noise known as speckle. The proposed system introduces the evaluation schema for segmentation of SAR images in Neutrosophic domain, which overcome the problem of speckle. The system is evaluated for different Swarm Optimization algorithms and the promoting results are obtained through the proposed evaluation schema. The results are discussed on the basis of different parameters.

Real-time Hand Tracking for Dynamic Gesture Recognition

Abstract: In the recent year, gesture recognition has become the most intuitive and effective communication technique for human interaction with machines. In this paper, we are going to work on hand gesture recognition and interpret the meaning of it from video sequences. Our work takes place in the following three phases: (1) hand detection and tracking, (2) feature extraction, and (3) gesture recognition. We have started proposed work with first step as applying hand tracking and hand detection algorithm to track hand motion and to extract position of the hand. Trajectory-based features are being drawn out from hand and used for recognition process, and hidden Markov model is being designed for each gesture for gesture recognition. Hidden Markov Model is basically a powerful statistical tool to model generative sequences. Our method is being tested on our own data set of 16 gestures, and the average recognition rate we have got is 91 %. With proposed methodology gives the better recognition results compare with the traditional approaches such as PCA, ANN, SVM, and DTW.

Face Veins Based MCMT Technique for Personal Identification

Abstract: Face veins based personal identification is a challenging task in the field of identity verification of a person. It is because many other techniques are not identifying the uniqueness of a person in the universe. This research paper finds the uniqueness of a person on the basis of face veins based technique. In this paper five different persons face veins images have been used with different rotation angles (left/right 90 to 270 and 315). For each person, eight different images at different rotations were used and for each of these images the same minimum cost minutiae tree (MCMT) is obtained. Here, Prim‟s or Kruskal‟s algorithm is used for finding the MCMT from a minutiae graph. The MCMT is traversed in pre-order to generate the unique string of vertices and edge lengths. We deviated the edge lengths of each MCMT by five pixels in positive and negative directions for robustness testing. It is observed in our experiments that the traversed string which consists of vertices and edge lengths of MCMT is unique for each person and this unique sequence is correctly identifying a person with an accuracy of above 95%. Further, we have compared the performance of our proposed technique with other standard techniques and it is observed that the proposed technique is giving the promising result. Indexed TermsFace Vein, Minimum Cost Minutiae Tree, Minutiae Pattern, Thermal Image, Minutiae Tree Traversal.

A novel hybrid intelligence approach for 2D packing through internet crowdsourcing

Abstract: Packing problems on its current state are being utilized for wide area of industrial applications. The aim of present research is to create and implement an intelligent system that tackles the problem of 2D packing of objects inside a 2D container, such that objects do not overlap and the container area is to be maximized. The packing problem becomes easier, when regular/rectangular objects and container are used. In most of the practical situations, the usage of irregular objects comes to existence. To solve the packing problem of irregular objects inside a rectangular container, a hybrid intelligence approach is introduced in our proposed work. The combination of machine intelligence and human intelligence is referred as the hybrid intelligence or semi-automated approach in the proposed methodology. The incorporation of human intelligence in the outcome of machine intelligence is possible to obtain using the internet crowdsourcing as we wish to handle the packing problem through internet crowdsourcing involving rural people. The proposed methodology is tested on different standard data sets and it is observed that it has clear advantage over both manual as well as fully automated heuristic based methods in terms of time and space efficiency.

Web-based Volume Visualization of 3D Medical Data using Slice Streaming Method

Abstract: In the advancement of Computer graphics tools/libraries used on Web-based services with its associated technologies and multimedia for making user friendly system. In this paper, we have proposed an approach for displaying the 3D Medical volume data visualization on the web-based system. Through which, the 3D volume dataset of many resolution sizes can be visualized onto the web using webGL with HTML5. Here we have used VTK tool for slicing the 3D Medical dataset. In the addition, this system can be accessed by many users at the same time. The 3D volume dataset is visualized with the user's perspective. If he/she wants to select any desired slice then the user can access and fetch that slice. This system has been tested by four different 3D Medical volume datasets with different resolution sizes from 6 MB to 3.15 GB.

Reduction of sample impoverishment problem in particle filter for object tracking

Abstract: Object tracking is one of the promising fields of research in the domain of image processing and computer vision. This paper deals with detection and tracking of moving objects. Detecting the region which is moving in the search region of a video is the first step. Tracking the object over frames is the second step and the most important one too. The first step can be achieved by background subtraction. Particle filter is used for estimating the object. In the second process, prediction and correction are the two tasks to achieve. In the process of prediction, resampling of particles in the search space region generates a sample impoverishment problem. To overcome this problem Continuous Opinion Dynamic Optimizer algorithm is used. This algorithm provides social rank to every particle for selection. CODO is combined with the particle filter algorithm and tested on the dataset. It contains different movements of objects randomly, partially occluded. The results of the proposed method are compared by calculating error difference from the previous method.

Empirical Study on Traffic Simulation Model in the Perspective of Raipur City

Abstract: Road network management is an important task for the economic and social growth. In the context of an Indian city, efficient road network management is required to overcome the problem of traffic congestions and road accidents. Computer traffic simulation is a better way to analysis the road network. In this study a maiden attempt for a systematic empirical traffic simulation study is performed in the perspective of Raipur city in India. The aim of this study is to set up, validate and experiment a simulation model of traffic by using Simulation of Urban Mobility (SUMO). General Terms Computer Simulation, Road Network Management, Modelling.