A voting-based technique for word spotting in handwritten document images

Multimedia Tools and Applications - Word spotting in handwritten document images is a field of immense interest due to its widespread applications. Recognition-free and recognition-based approaches...     

Toward automating Hammersmith pulled-to-sit examination of infants using feature point based video object tracking

Hammersmith Infant Neurological Examination (HINE) is a set of tests used for grading neurological development of infants on a scale of 0 to 3. These tests help in assessing neurophysiological development of babies, especially preterm infants who are born before (the fetus reaches) the gestational age of 36 weeks. Such tests are often conducted in the follow-up clinics of hospitals for grading infants with suspected disabilities. Assessment based on HINE depends on the expertise of the physicians involved in conducting the examinations. It has been noted that some of these tests, especially pulled-to-sit and lateral tilting, are difficult to assess solely based on visual observation. For example, during the pulled-to-sit examination, the examiner needs to observe the relative movement of the head with respect to torso while pulling the infant by holding wrists. The examiner may find it difficult to follow the head movement from the coronal view. Video object tracking based automatic or semi-automatic analysis can be helpful in this case. In this paper, we present a video based method to automate the analysis of pulled-to-sit examination. In this context, a dynamic programming and node pruning based efficient video object tracking algorithm has been proposed. Pulled-to-sit event detection is handled by the proposed tracking algorithm that uses a 2-D geometric model of the scene. The algorithm has been tested with normal as well as marker based videos of the examination recorded at the neuro-development clinic of the SSKM Hospital, Kolkata, India. It is found that the proposed algorithm is capable of estimating the pulled-to-sit score with sensitivity (80%-92%) and specificity (89%-96%).     

Matrices over semirings

Introducing the concept of semi-invertibility of square matrices over semirings, some sufficient conditions for semi-invertibility of square matrices over various kinds of semirings are obtained. Also a necessary and sufficient condition for semi-invertibility of square matrices over semirings is obtained.     

Synthesis and characterization of indium xanthates and their use for the preparation of beta-In2S3 nanoparticles

Synthesis and characterization of various classical indium xanthate complexes of the type [InCl(3-n)(S2COR)n] (n = 1, 2, or 3; R = Me, Et, Pr(i), and Bu(s)) have been discussed. Crystalline beta-ln2S3 nanoparticles were obtained by the solvent thermolysis of indium tris-alkylxanthates in ethylene glycol at 196 degrees C, and were characterized by elemental analysis, IR, powder XRD, and XPS techniques. TEM results showed that the size of beta-In2S3 nanoparticles depended on the nature of the precursor used. The optical properties of beta-In2S3 nanocrystals have shown quantum confinement of the excitonic transition.     

Database intrusion detection using sequence alignment

Information is considered to be the most valuable asset of any organization and hence, it should be securely maintained. However, rapid proliferation of the Internet and Web applications has increased the threat of information security breaches. Traditional database security mechanisms are often not sufficient to protect sensitive information against novel attacks. Intrusion detection systems (IDS) are used to detect any such intrusion, once traditional security mechanisms have been compromised. User-level profile is effective for database intrusion detection, but maintaining such profiles is not practical for an organization with a large number of users. Thus, an IDS needs to be flexible enough to choose a profile granularity according to the type of the organization. Further, only intra-transactional pattern matching for intrusion detection is not quite effective for detecting intrusion in a database. We propose an IDS that uses inter-transactional as well as intra-transactional features for intrusion detection. It supports selection of profile and transactional feature granularity as well. We use sequence alignment as a tool for comparing database access patterns of genuine users and intruders.     

State-based modeling and object extraction from echocardiogram video

In this paper, we propose a hierarchical state-based model for representing an echocardiogram video. It captures the semantics of video segments from dynamic characteristics of objects present in each segment. Our objective is to provide an effective method for segmenting an echo video into view, state, and substate levels. This is motivated by the need for building efficient indexing tools to support better content management. The modeling is done using four different views, namely, short axis, long axis, apical four chamber, and apical two chamber. For view classification, an artificial neural network is trained with the histogram of a region of interest of each video frame. Object states are detected with the help of synthetic M-mode images. In contrast to traditional single M-mode, we present a novel approach named sweep M-mode for state detection. We also introduce radial M-mode for substate identification from color flow Doppler 2-D imaging. The video model described here represents the semantics of video segments using first-order predicates. Suitable operators have been defined for querying the segments. We have carried out experiments on 20 echo videos and compared the results with manual annotation done by two experts. View classification accuracy is 97.19%. Misclassification error of the state detection stage is less than 13%, which is within acceptable range since only frames at the state boundaries are found to be misclassified.     

Broadband scalable model for Si‐RF on‐chip spiral inductors with substrate eddy current effect

A new broadband scalable compact model for on‐chip spiral inductor is presented. It includes a substrate network and a high resistive component to accurately capture the substrate eddy current effect, skin and proximity effects at high frequencies. The proposed model shows a good match with the measurement results over a frequency range of up to 20 GHz. An automated parameter extraction and optimization procedure has been developed using equivalent circuit analysis and particle swarm optimization. We achieved good scalability for all model parameters with respect to spiral layout geometry. © 2007 Wiley Periodicals, Inc. Int J RF and Microwave CAE, 2007.     

Occlusion detection and gait silhouette reconstruction from degraded scenes

Gait, which is defined as the style of walking of a person, has been recognized as a potential biometric feature for identifying human beings. The fundamental nature of gait biometric of being unconstrained and captured often without a subject’s knowledge or co-operation has motivated many researchers over the last one decade. However, all of the approaches found in the literature assume that there is little or no occlusion present at the time of capturing gait images, both during training and during testing and deployment. We look into this challenging problem of gait recognition in the presence of occlusion. A novel approach is proposed, which first detects the presence of occlusion and accordingly extracts clean and unclean gait cycles from the whole input sequence. In the second step, occluded silhouette frames are reconstructed using Balanced Gaussian Process Dynamical Model (BGPDM). We evaluated our approach on a new data set TUM-IITKGP featuring inter-object occlusion. Algorithms have also been tested on CMU’s Mobo data set by introducing synthetic occlusion of different degrees. The proposed approach shows promising result on both the data sets.     

Alzheimer's detection using various feature extraction approaches using a multimodal multi-class deep learning model

Alzheimer's disease is a chronic brain condition that takes a toll on memory and potential to do even the most basic tasks. With no specific solution viable at this time, it's critical to pinpoint the start of Alzheimer's disease so that necessary steps may be initiated to limit its progression. We used three distinct neuroanatomical computational methodologies namely 3D-Subject, 3D-Patches, and 3D-Slices to construct a multimodal multi-class deep learning model for three class and two class Alzheimer's classification using T1w-MRI and AV-45 PET scans obtained from ADNI database. Further, patches of various sizes were created using the patch-extraction algorithm designed with torch package leading to separate datasets of patch size 32, 40, 48, 56, 64, 72, 80, and 88. In addition, Slices were produced from images using either uniform slicing, subset slicing, or interpolation zoom approaches then joined back to form a 3D image of varying depth (8,16,24,32,40,48,56, and 64) for the Slice-based technique. Using T1w-MRI and AV45-PET scans, our multimodal multi-class Ensembled Volumetric ConvNet framework obtained 93.01% accuracy for AD versus NC versus MCI (highest accuracy achieved using multi-modalities as per our knowledge). The 3D-Subject-based neuroanatomy computation approach achieved 93.01% classification accuracy and it overruled Patch-based approach which achieved 89.55% accuracy and Slice-Based approach that achieved 89.37% accuracy. Using a 3D-Patch-based feature extraction technique, it was discovered that patches of greater size (80, 88) had accuracy over 89%, while medium-sized patches (56, 64, and 72) had accuracy ranging from 83 to 88%, and small-sized patches (32, 40, and 48) had the least accuracy ranging from 57 to 80%. From the three independent algorithms created for 3D-Slice-based neuroanatomy computational approach, the interpolation zoom technique outperformed uniform slicing and subset slicing, obtaining 89.37% accuracy over 88.35% and 82.83%, respectively. Link to GitHub code: https://github.com/ngoenka04/Alzheimer-Detection .