Accepted Papers

  • Gesture Recognition of American Sign Language with Prime Sense Sensors using MATLAB
    Manjusha Choorakuzil,Unnikrishnan Parameswaran and Micheal Buckley ,University at Buffalo,United States.
    ABSTRACT
    Sign Language is the primary medium of communication for the hearing impaired and mute in our society. Like any good Sign Language, American Sign Language (ASL) does justice to its job in expressing flow of thoughts and emotions as good as any spoken language. It involves the motion of hands and body posture as well as facial expressions. This paper is an extension to the work done in our previous paper where we classified words in Indian Sign Language [1]. The primary aim of this paper is to recognize dynamic gestures in American Sign Language and display the corresponding recognized words in text format. These recognized words could then be used by various applications in devices or be converted to a voice based output.With the advent of perceptual computing we are approaching a stage where touch based input would soon be replaced by motion sensors. With devices like Microsofts Kinect and Prime Sense sensors, the phrase Talk to the hand would come to life when it comes to devices. Of late, lot of research has been going on in the area of 3D geometric processing of sequence of images.In our project, we captured a sequence of 3D ASL gestures using the Prime Sense sensor and proposed a novel trajectory based method of classification of ASL gestures using the method of Axis of Least Inertia (ALI).Feature Extraction is done using both local features and global features of a gesture. The ALI method of feature extraction is applied to global features while local feature extraction is done using the distance between each of the fingertips to the centroid. Integrating the results of local and global recognition improved the classification accuracy and system performance. Other applications of the system include using the system as a sign language tutor and in cafeterias, ticket counters and other public places where systems could be employed to accept gesture input from the users to process the needs of a person with hearing disability.
  • Correlation Based Hindi Word Sense Disambiguation
    Madhavi Agarwal and Jyoti Bajpai,GLA UNIVERSITY,India
    ABSTRACT
    Today internet usage has seen tremendous growth. As English is the primary language, documents are mostly available in English language. In India, Hindi is the prevalent language and user wants to access data in Hindi. For the language processing we are required to get the exact sense of polysemous word interpreting the meaning in a particular context. To disambiguate the meaning of the polysemous word, the techniques used is Word Sense Disambiguation (WSD). It is a known problem in natural language processing referred as lexical semantic ambiguity. In this paper, correlation analysis of context in which the target word is used with the collocation vector of definition of target word derived from Hindi WordNet i.e. developed at IIT Bombay and the co-occurrence vector which is derived from Hindi Corpus is computed. The proposed approach uses collocation information, co-occurrence information of target word to assign weights to the different senses of ambiguous word. The evaluation is done on the 60 ambiguous words, precision obtained is 88.92%. The proposed experiment shows better efficiency.
  • Driver drowsiness monitoring based on eye and yawn detection
    Shubhangi Kalyane and Parmindar Kaur,J.N.E.C,India
    ABSTRACT
    Drowsiness can be dangerous when performing tasks that require constant concentration, such as driving a vehicle. When a person is sufficiently fatigued, drowsiness may be experienced. Drowsy driving is a prevalent and serious public health issue that deserves more attention, education,and policy initiatives so a substantial amount of lives can be saved and disability averted due to drowsy driving accidents.In an effort to reduce the number of fatigue-related crashes and to save lives, special body and face gestures are used as sign of driver fatigue, including yawning, and eye movement,which indicate that the driver is no longer in a proper driving condition. In this paper, we discuss a method for detecting driver’s drowsiness and subsequently alerting the driver as well as the owner of the vehicle.
  • Intelligent Adaptive Learning in a Changing Environment
    Quentin Berthelot and Guillaume Valentis,ECE Paris Graduate School of Engineering,France
    ABSTRACT
    Nowadays the trend to develop ever more intelligent and autonomous systems often takes its inspiration in the living beings on Earth. Some simple isolated systems are able, once brought together, to form a strong and reliable system. When trying to adapt the idea to man-made systems it is not possible to include in their program everything the system may encounter during its life cycle. It is thus necessary to make the system able to take decisions based on other criteria such as its past experience, ie to make the system learn on its own. However at some point the acquired knowledge depends also on environment. So the question is: if system environment is modified, how could the system respond to it quickly and appropriately enough? Here, starting from reinforcement learning to rate its decisions, and using adaptive learning algorithms for gain and loss reward, the system is made able to respond to changing environment and to adapt its knowledge as time passes. Application is made to a robot finding an exit in a labyrinth.
  • Detection Of Blood Vessels And Measurement Of Vessel Width For Diabetic Retinopathy
    Abinaya Seenivasan,Sukanya S and Saranya N,Thiagarajar College of Engineering,India
    ABSTRACT
    The proposed method measures the retinal blood vessel diameter to identify arteriolar narrowing,arteriovenous (AV) nicking, branching coefficients to detect early diabetic retinopathy. It utilizes the vessel centerline and edge information to measure the width for a vessel segment. From the input retinal image, the vascular network is extracted using the local entropy thresholding method. The vesselboundaries are extracted using sobel edge detection method. The skeletonization operation is applied to the vascular network and mapping the vessel boundaries and the skeleton image. The branching point detection method is then performed to localize all crossing locations. A rotational invariant mask to search the pixel pairs from the edge image, and calculate the shortest distance pair which provides the vessel width (or diameter) for that cross-section. Variation in the width measurement identifies the diabetic retinopathy.
  • Phoneme Recognition - Neural Network Implementation
    Sriniwas Surampudi and Ritu Pal,VNIT, India
    ABSTRACT
    This paper shows the utilization of neural network’s parallel characteristics as well as self learning characteristics in phoneme recognition. This paper demonstrates the utility of machine learning algorithms in signal processing. Different types of neural networks are applied at different stages of the whole process. Artificial neural network’s implementation has improved performance of feature extraction, and matching techniques of phoneme recognition.
  • A Belief Revision System for Logic Programs
    Taher Ali1, Mohd Sapiyan1and Ziad Najem2,1Gulf University for Science and Technology,Kuwait and 2Kuwait University,Kuwait.
    ABSTRACT
    Search is one of the most important needs of problem solvers. Usually the problem solvers suffer from retracing conclusions. If a problem solver cached its inference, then it would not need to retrace conclusions that it had already derived earlier in the search. By caching the inferences, the problem solver avoid throwing away useful results and avoid wasting effort rediscovering the same things over and over. In this paper we present a belief revision system for logic programs that can work under the non-monotonic logic.