Search

This paper presents a new image fusion algorithm based on the visible and thermal IR face images, which exploits the advantages of both kinds of images. The proposed fusion algorithm utilizes the translation-invariant à-trous wavelet transform and random forest (RF) classifier to decide the contribution of the visible and thermal IR face images in the formation of fused images. The Universal Image Quality Index is used to evaluate the quality of fused images and results are quite satisfactory. The fused face images are recognized by RF classifiers. The recognition performances of the proposed fusion scheme are 99.07% and 100% for UGC-JU and IRIS benchmark face databases respectively, which is better than those if only visible or thermal IR faces are used. [ABSTRACT FROM AUTHOR]

The paper mainly aims to create a mapping between facial expression and its corresponding facial muscle contractions along with their movement directions. This mapping is illustrated in terms of mathematical symbolic representations. The paper proposes a set of mathematical representations of basic as well as blended facial expressions. These symbolic representations of facial expressions are evaluated from different normalized facial features and 2D spatial coordinates of a face. They offer a simple generalization of 18 facial expressions and will be used as a background formulation for generating expressive face image from a given neutral face. The facial expression modeling and synthesis is a widely useful application for man-machine interaction. [ABSTRACT FROM AUTHOR]

On human face, non-rigid facial movements due to facial expressions cause noticeable alterations in their usual shapes, which sometimes create occlusions in facial feature areas making face recognition as a difficult problem. The paper presents an automatic Region of Interest (ROI) detection technique of six universal expressive face images. The proposed technique is a facial geometric based hybrid approach. The localization accuracy was evaluated by rectangular error measure and was tested on Japanese Female Facial Expression (JAFFE) database. The average localization accuracy of all detected facial regions is 94%. [ABSTRACT FROM AUTHOR]

Face is the most significant biometric as it reveals a person's identity more accurately. Soft biometric traits like facial marks have played a crucial role in identifying a human face. The paper presents an automatic prominent mole detection and validation technique which can reduce the adverse effect of illumination in face recognition. Normalized cross-correlation with LoG filter is used to detect the facial mole candidates. A contributory threshold based step is introduced in this paper to improve the accuracy of the mole detector. The mole detection rate is 91.67% using our own developed “DeitY-TU” face database and 90.58% using FEI face database. [ABSTRACT FROM AUTHOR]

Abstract: In academia Plagiarism means copying of other work without author's permission. Presently available system mainly focuses on software plagiarism. They mainly based on token analysis, linguistic patterns, taxonomy and textual features. In this paper we mainly concentrate on research papers to check whether the documents are plagiarized or not. So far not much work has been done to detect plagiarism in research document. Our work focuses on the similarity of different simple equations present in a document. It can easily extract those equations from the documents, compare them even if the variables are changed in plagiarized document with the original one and can detect if the document is plagiarized or not. This method will not work if the research paper does not contain any equation. [Copyright &y& Elsevier]

Abstract: A region based multi-sensor image fusion approach is proposed in this paper. At the initial stage of our algorithm, noise is suppressed from the input images by applying a 3 × 3 filter mask. In the next phase, regions are segmented from the input images by computing similarity map image followed by marker based watershed algorithm. Thereafter, regions are fused by computing the relative importance of a pixel in the region. Here, the relative importance of a pixel in the region is calculated as the second central moment of that pixel in the neighborhood with respect to the asymmetry or skewness of the whole region. After that a decision map is implemented based on the relative importance of a pixel in the region for fusion of the two correspondence regions. Finally, all the fused regions are combined to produce a final fused image. To check the robustness of our algorithm, we have tested it on 120 multi-sensor image pairs collected from Manchester University UK database and compared with some state-of-the-art region based fusion techniques. The experimental result shows the superiority of our proposed method in terms of visual and objective perception evaluation indexes. [Copyright &y& Elsevier]

Abstract: In this paper, a thermal infra red face recognition system for human identification and verification using blood perfusion data and back propagation feed forward neural network is proposed. The system consists of three steps. At the very first step face region is cropped from the colour 24-bit input images. Secondly face features are extracted from the croped region, which will be taken as the input of the back propagation feed forward neural network in the third step and classification and recognition is carried out. The proposed approaches are tested on a number of human thermal infra red face images created at our own laboratory. Experimental results reveal the higher degree performance. [Copyright &y& Elsevier]

• Design a Completely Automated Public Turing Test to Tell Computers & Humans Apart. • Human verification is proposed to thwart the probability of attacks. • Spoofing attack detection using image quality assessment is addressed. • Finger geometry and a feature selection algorithm is proposed. • Enhanced security and accuracies are achieved. This paper presents a human verification scheme in two independent stages to overcome the vulnerabilities of attacks and to enhance security. At the first stage, a hand image-based CAPTCHA (HandCAPTCHA) is tested to avert automated bot-attacks on the subsequent biometric stage. In the next stage, finger biometric verification of a legitimate user is performed with presentation attack detection (PAD) using the real hand images of the person who has passed a random HandCAPTCHA challenge. The electronic screen-based PAD is tested using image quality metrics. After this spoofing detection, geometric features are extracted from the four fingers (excluding the thumb) of real users. A modified forward–backward (M-FoBa) algorithm is devised to select relevant features for biometric authentication. The experiments are performed on the Boğaziçi University (BU) and the IIT-Delhi (IITD) hand databases using the k -nearest neighbor and random forest classifiers. The average accuracy of the correct HandCAPTCHA solution is 98.5%, and the false accept rate of a bot is 1.23%. The PAD is tested on 255 subjects of BU, and the best average error is 0%. The finger biometric identification accuracy of 98% and an equal error rate (EER) of 6.5% have been achieved for 500 subjects of the BU. For 200 subjects of the IITD, 99.5% identification accuracy, and 5.18% EER are obtained. [ABSTRACT FROM AUTHOR]

microRNAs (miRNAs) are short, non-coding, endogenous RNA molecule that regulates messenger RNAs (mRNAs) at the post-transcriptional level. The discovery of this regulatory relationship between miRNAs and mRNAs is an important research direction. In this regard, our method proposed an integrated approach to identify the mRNA targets of dysregulated miRNAs using next-generation sequencing data from six cancer types. For this analysis, a sensible combination of data mining tools is chosen. In particular, Random Forest, log-transformed Fold change, and Pearson correlation coefficient are considered to find the potential miRNA–mRNA pairs. During this study, we have identified six cancer-specific overlapping sets of miRNAs whose classification accuracy is always higher than 91%. Furthermore, a promising correlation signature of significantly dysregulated miRNAs and mRNAs are recognized. A comprehensive analysis found that the cumulative percentage of negative correlation coefficients is higher than its positive counterpart. Moreover, experimentally validated miRNA–target interactions databases called miRTarBase is used to validate significantly correlated mRNAs. According to our study, the smallest set of significantly dysregulated miRNAs is 43 in PRAD data, while for mRNAs the smallest set is 238 in the LUAD cancer type. The obtained miRNA–mRNA pairs are subjected to do pathway enrichment analysis and gene ontology analysis. Regulatory roles of these dysregulated miRNAs with associated diseases are identified by constructing a regulatory network between miRNAs and associated diseases. Moreover, the relation between miRNAs expression level and patient survival is also analyzed. To conclude, the miRNA–mRNA pairs identified in this study may serve as promising candidates for subsequent in-vitro validation. [ABSTRACT FROM AUTHOR]

• Measure of Inflammation of joints in arthritis patients is important to the clinicians for the treatment of arthritis. • Pathological and clinical examination often fail to detect subclinical inflammation. • Thermal imaging is able to detect the subclinical inflammation. • Inflamed areas are represented as a hotspot in thermal images of the arthritis patients with inflammation. • To overcome under segmentation two under segmented results can be intersected. Original image and its different features such as entropy, energy can be used to obtain two under segmented results. In arthritis, subclinical inflammation referred to the clinical condition when rheumatologists are in confusion about the presence of inflammation using clinical and pathological observations. Application of Thermal imaging in detection of subclinical inflammation is highlighted in this literature. Segmentation of the hotspot area from the thermal image is the initial step for further analysis of the hotspot. Analysis of the hotspot will help in prediction of the subclinical inflammation, impact of inflammation. Methodologies reported in existing literature for segmentation of hotspot or inflamed knee region in medical thermal images suffer from over and under extraction. In the present scope, we try to overcome this limitation by extending the conventional region growing segmentation technique with stronger similarity criteria and stopping rule. In this method, hotspot or inflamed region is generated by taking the intersection of two independent regions produced by two different version of Region growing algorithm using a separate set of parameters. An automatic multiseed selection procedure ensures prevention of missed segmentation. We validate our technique by experimentation on various thermal image datasets like a newly created inflammatory thermal knee-joint-Database of 50 images, DBT-TU-JU Dataset, and DMR-IR Dataset. The effectiveness of the proposed technique is established compared to the performance of state-of-the-art competing methodologies. [ABSTRACT FROM AUTHOR]

Highlights • In this work, we have developed a patch-based classifier (PBC) using the convolutional neural network (CNN) for automated classification of breast cancer histopathology images into 4-different histology class namely normal, benign, in situ and invasive carcinoma. • The developed patch-based classifier (PBC) uses an optimal architecture of a convolutional neural network (CNN), for automated classification of breast cancer histopathology images. • The proposed classification system works in two different modes: one patch in one decision (OPOD) and all patches in one decision (APOD). The patch labels are predicted by OPOD mode, and the result is obtained unanimously whereas in the APOD mode class label of the image is obtained by a majority voting scheme. • To verify the classification ability of the proposed system, the breast histopathological images are classified into 2 classes (non-malignant and malignant) as well as 4 classes (normal, benign, in situ and invasive carcinoma) while most of the existing methods classify the same broadly into 2 classes. • We have also explored the potentiality of our proposed model in classifying the images in the test dataset obtained by splitting the training set as well as the actual hidden test dataset of ICIAR-2018 breast cancer histology image dataset. • Our model achieves an accuracy of 87% in classifying the images of ICIAR-2018 hidden test dataset. Abstract In this work, we proposed a patch-based classifier (PBC) using Convolutional neural network (CNN) for automatic classification of histopathological breast images. Presence of limited images necessitated extraction of patches and augmentation to boost the number of training samples. Thus patches of suitable sizes carrying crucial diagnostic information were extracted from the original images. The proposed classification system works in two different modes: one patch in one decision (OPOD) and all patches in one decision (APOD). The proposed PBC first predicts the class label of each patch by OPOD mode. If that class label is the same for all the extracted patches and that is the class label of that image, then the output is considered as correct classification. In another mode that is APOD, the class label of each extracted patch is extracted as done in OPOD and a majority voting scheme takes the final decision about class label of the image. We have used ICIAR 2018 breast histology image dataset for this work which comprises of 4 different classes namely normal, benign, in situ and invasive carcinoma. Experimental results show that our proposed OPOD mode achieved a patch-wise classification accuracy of 77.4% for 4 and 84.7% for 2 histopathological classes respectively on the test set obtained by splitting the training dataset. Also, our proposed APOD technique achieved image-wise classification accuracy of 90% for 4-class and 92.5% for 2-class classification respectively on the split test set. Further, we have achieved accuracy of 87% on the hidden test dataset of ICIAR-2018. [ABSTRACT FROM AUTHOR]

This paper presents a novel approach for Human Face Recognition, namely Regularized Bi-partitioned Entropy Component Analysis (RBECA). This conservative approach regularizes the kernel entropy components by deterring the noise and affecting the lower entropy regions area, making the method robust to noise. The kernel feature space, formed by the kernel entropy component analysis (KECA), is divided into two partitions: the High Entropy Space (HES) and the Low Entropy Space (LES). The noise-laden low entropy spectrum is regularized by predicting entropy values obtained from the information-filled High Entropy Spectrum. The corresponding projection vectors are adjusted accordingly. A null space, comprising the negligible information and many dimensions, is eliminated using a Golden Search minimization function at two stages. The method retains the maximum entropy property and high recognition accuracy while using the optimum number of features. This resultant feature vector is classified using the cosine similarity measure. The algorithm is successfully tested on several benchmark databases like AR, FERET, FRAV2D, and LFW, using standard protocols and compared with other competitive methods. The proposed method achieves much better recognition accuracy than other well-known methods like PCA, ICA, KPCA, KECA, LGBP, ERE, etc., in all considered cases. Moreover, we have also proposed a CNN for the comparative analysis. For unbiased or fair performance evaluation, the sensitivity and specificity are also reported. • A noise robust RBECA method is proposed for human face recognition. • The proposed method requires a lower number of features than KECA. • The highest level of discriminatory information is retained. • The algorithm is trained and tested on several benchmark face databases. • A deep learning CNN framework is also implemented for comparative analysis. [ABSTRACT FROM AUTHOR]

The proposed work describes an efficient methodology for adaptive background modeling and foreground extraction from video data using the newly proposed Spatio-temporal region persistence (STRP) descriptor. The STRP background descriptor includes block-wise statistics of intensity bins and their temporal persistency. Blockwise feature extraction helps to consider the local changes, while intensity bins provide consistent output for a group of similar intensities in an intra-regional sense. In this work, we have tried to minimize the effect of different video irregularities like dynamic background, ghosting effect, change in illuminations, video noise, etc. Additionally, adaptive threshold selection and regular adjustment of modeled background descriptors make the procedure robust. Two benchmark datasets, Changed Detection and Scene Background Modeling, and Initialization (SBMI) have been used to verify the efficiency of our work. The results and comparative studies with the related works justify the effectiveness of our proposed technique. [ABSTRACT FROM AUTHOR]

Highlights • The proposed method works on face recognition as well as facial expression recognition system. • The method recognizes noisy face images also. • Log-ICA II performs better than Log-ICA I. Abstract By reducing the gaussianity, Independent Component Analysis (ICA) behaves robustly in segregating individual signals of non-skewed characteristic from a mixed composite signal. In this article, we present a next-generation variant of ICA, especially applicable in the skewed composite signal scenario, applying the Logarithmic transformation on basic ICA, named as Log-ICA. This approach is capable of decreasing overlapping probability densities of the composite signal, which, in turn, extracts more independent components because of reduced gaussianity. Here also we use two different architectures Log-ICA I and Log-ICA II corresponding to two variants of ICA architecture (ICA I and ICA II). We justify the effectiveness of the proposed technique on five separate benchmark face datasets using five classifiers. Out of five face datasets, two datasets contain both visible and thermal face images. Experimental results show that Log-ICA II performs better than Log-ICA I and two variants of ICA for original face images and noise-induced face images. [ABSTRACT FROM AUTHOR]