Landslide Susceptibility Mapping through Weightages Derived from Statistical Information Value Model
An Efficient Foot Ulcer Determination System for Diabetic Patients
Statistical Wavelet based Adaptive Noise Filtering Technique for MRI Modality
Real Time Sign Language: A Review
Remote Sensing Schemes Mingled with Information and Communication Technologies (ICTS) for Flood Disaster Management
FPGA Implementation of Shearlet Transform Based Invisible Image Watermarking Algorithm
A Comprehensive Study on Different Pattern Recognition Techniques
User Authentication and Identification Using NeuralNetwork
Flexible Generalized Mixture Model Cluster Analysis withElliptically-Contoured Distributions
Efficient Detection of Suspected areas in Mammographic Breast Cancer Images
Emotion recognition remains a potential research area as efforts to make machines to mimic humans in most areas of human life is yet actualized. This paper presents an emotion recognition system, to enhance recognition accuracy for better user experience. Principal Component Analysis (PCA) was implemented via Singular Value Decomposition (SVD) and used for feature extraction process. In classification process, Discrete Hidden Markov Model (HMM) was utilized in a principled manner. Two-dimensional spatial face features were realized by varying quantization levels. The quantization level with the efficient feature description, judged by the highest recognition accuracy was chosen to train the system. The recognition accuracy of the system was studied on two publicly available datasets, namely, JAFFE and Cohn Kanade (CK) datasets. The system showed better performances compared with other state of the art systems.
This paper presents the development of a new optimization algorithm called the Smell Agent Optimization (SAO). The algorithm uses the phenomenon of smell and the intuitive trailing behavior of an agent to identify a smell source. The developed algorithm has two basic modes used in the optimization process, which are the sniffing mode and trailing mode. In the sniffing mode, the evaporation of smell molecules from a source is modeled and in the trailing mode, the movement of an agent towards the smell molecules is modeled. The performance of SOA was evaluated using 10 benchmark functions and results was compared with PSO, ABC, and GA. Simulation results showed the efficiency of the developed SAO in solving unimodal and multimodal functions.
In recent times, big Internet companies have come under increased pressure from governments and NGOs to remove inappropriate materials from social media platforms (e.g., Twitter, Facebook, YouTube). A typical example of this problem is the posting of hateful, abusive, and violent tweets on Twitter which has been blamed for inciting hatred, violence and causing societal disturbances. Manual identification of such tweets and the people who post these tweets is very difficult because of the large number of active users and the frequency with which such tweets are posted. Existing approaches for identifying inappropriate tweets have focused on the detection of such tweets without identifying the users who post them. This paper proposes an approach that can automatically identify different types of inappropriate tweets together with the users who post them. The proposed approach is based on a user profiling algorithm that uses a deep Long Short-Term Memory (LSTM) based neural network trained to detect abusive language. With the support of word embedding features learned from the training set, the algorithm is able to classify the tweets of users into different abusive language categories. Thereafter, the user profiling algorithm uses the classes assigned to the tweets of each user to profile each user into different abusive language category. Experiments on the test set show that the deep LSTM-based abusive language detection model reached an accuracy of 89.14% on detecting whether a tweet is bigotry, offensive, racist, extremism-related and neutral. Also, the user profiling algorithm obtained an accuracy of 83.33% in predicting whether a user is a bigot, racist, extremist, uses offensive language and neutral.
Diabetes Mellitus (DM) and Thyroid are the major coexistent autoimmune disorders affecting people globally. Due to prolonged chronic mental stress in the modern lifestyle, Thyroid disorder is affecting all age groups and people with Thyroid disorder have an increased risk of developing DM complications. Because, abnormal Thyroid dysfunction can have dreadful effects on blood glucose control and can affect the course of DM. This paper proposes a Medical Expert system to assist clinicians in predicting the prevalence of developing autoimmune DM more precisely in patients suffering from Thyroid and further helps to investigate better in the line of improving public health. In this work, Fuzzy logic based inference system and unsupervised machine learning algorithms are used to discover associations and dependencies between Thyroid and DM. The inferred knowledge base is used to design Fuzzy based Expert system. To develop a more realistic expert system, blood sample reports of people affected by DM and Thyroid disorder have been collected from various Endocrine centres in Andhra Pradesh, India. Specificity, Sensitivity, Predictive Values, and Likelihood Ratios of the proposed system are promising in support of system functionality
This paper attempts to decipher old documents using symbol to script mapping scheme. Symbols are confined to documents either as isolated notations or handwritten texts with a number of not able features. This paper describes a method to separate and classify handwritten non-cursive symbols in Grantha script. This work uses statistical correlation coefficient method for separation and classification, without the recognition of the symbols. The Grantha script symbols mapping model comprises of selection, separation, preprocessing, classification, and finally mapping. The proposed model employs bounding box algorithm for locating the symbols. The algorithm selects the symbols and excludes the non-symbol components to an extent possible. For experiments, 135 Grantha script document images of varying deteriorating complexities were used. The resulting symbol classification rate (i.e., the proportion of symbols automatically classified) was obtained near to 80%, aiding in mapping to a predetermined mapping scheme.
