Department of Computer Science & Engineering

&ldquo;IIT Jodhpur is committed to impart quality education to facilitate the creation of qualified, competent and responsible human resources to meet the emerging technological challenges of the world. It is working towards creating an equitable learning platform that is positive, accessible and effective in the face of the modern societal and scientific advancements.

Flexibility in curriculum, in all its academic programs, encompassing elements of teaching, research, creativity, innovation and entrepreneurship is the most unique feature of IIT Jodhpur. The Department of Science and Technology (DST), under the National Mission on Cyber Physical Systems, has identified IIT Jodhpur as a Technology &amp; Innovation Hub in the area of Augmented Reality and Virtual Reality. The institute has also been nominated as the coordinator for the Jodhpur City and Knowledge Cluster by the Office of Principal Scientific Adviser, Government of India.

The sprawling 852-acre IIT Jodhpur campus is a majestic tribute to the deep-rooted diverse cultural heritage of the region. The campus is an architectural masterpiece, which is slated to become an international exemplar of sustainability with net-zero energy, and advanced water and waste management strategies.&rdquo;

IIT Jodhpur

Covolutional neural networks extract deep features from input image. The features are invariant to small distortions in the input, but are sensitive to rotations, which makes them inefficient to classify rotated images. We propose an architecture that requires training with images having digits at one orientation, but is able to classify rotated digits oriented at any angle. Our network is built such that it uses any simple unit of CNN by training it with single orientation images and uses it multiple times in testing to accomplish rotation invariant classification. By using CNNs trained with prominent features of images, we create a stacked architecture which gives adequately satisfactory classification accuracy. We demonstrate the architecture on handwritten digit classification and on the benchmark mnist-rot-12k. The introduced method is capable of roughly identifying the orientation of digit in an image. © 2017, Springer International Publishing AG.

World University root node

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

Stacked Features Based CNN for Rotation Invariant Digit Classification

Background Effective connectivity (EC) analysis of neuronal groups using fMRI delivers insights about functional-integration. However, fMRI signal has low-temporal resolution due to down-sampling and indirectly measures underlying neuronal activity. New method The aim is to address above issues for more reliable EC estimates. This paper proposes use of autoregressive hidden Markov model with missing data (AR-HMM-md) in dynamically multi-linked (DML) framework for learning EC using multiple fMRI time series. In our recent work (Dang et al., 2016), we have shown how AR-HMM-md for modelling single fMRI time series outperforms the existing methods. AR-HMM-md models unobserved neuronal activity and lost data over time as variables and estimates their values by joint optimization given fMRI observation sequence. Results The effectiveness in learning EC is shown using simulated experiments. Also the effects of sampling and noise are studied on EC. Moreover, classification-experiments are performed for Attention-Deficit/Hyperactivity Disorder subjects and age-matched controls for performance evaluation of real data. Using Bayesian model selection, we see that the proposed model converged to higher log-likelihood and demonstrated that group-classification can be performed with higher cross-validation accuracy of above 94% using distinctive network EC which characterizes patients vs. controls. The full data EC obtained from DML-AR-HMM-md is more consistent with previous literature than the classical multivariate Granger causality method. Comparison The proposed architecture leads to reliable estimates of EC than the existing latent models. Conclusions This framework overcomes the disadvantage of low-temporal resolution and improves cross-validation accuracy significantly due to presence of missing data variables and autoregressive process. © 2016 Elsevier B.V.

School of Artificial Intelligence and Data Science (AIDE)

Journal of Neuroscience Methods

Learning effective connectivity from fMRI using autoregressive hidden Markov model with missing data

Cross-spectral cross-resolution face recognition in videos

CAPTCHAs are an important security tool for preventing automated attacks against online systems. However, they can be an undesirable source of friction in the user experience. In this poster, we propose a novel approach to dynamically improve the CAPTCHA experience for users by individually learning which types of CAPTCHA tests each user is best at solving and adapting future tests presented based on this knowledge.

system

Poster: Adaptcha: An Adaptive CAPTCHA for Improved User Experience

For a family of graphs (formula presented), the (formula presented) -Contraction problem takes as an input a graph G and an integer k, and the goal is to decide whether there exists (formula presented) of size at most k such that G/F belongs to (formula presented). When (formula presented) is the family of paths, trees or cacti, then the correspond-ing problems are Path Contraction, Tree Contraction and Cac-tus Contraction, respectively. It is known that Tree Contraction and Cactus Contraction do not admit a polynomial kernel unless (formula presented), while Path Contraction admits a kernel with O(k) vertices. The starting point of this article are the following natural ques-tions: What is the structure of the family of paths that allows Path Con-traction to admit a polynomial kernel? Apart from the size of the solu-tion, what other additional parameters should we consider so that we can design polynomial kernels for these basic contraction problems? With the goal of designing polynomial kernels, we consider the family of trees with bounded number of leaves (note that the family of paths are trees with at most two leaves). In particular, we study Bounded Tree Contraction (Bounded TC). Here, an input is a graph G, integers k and, (formula presented) and the goal is to decide whether or not, there exists a subset (formula presented) of size at most k such that G/F is a tree with at most leaves. We design a kernel for Bounded TC with O(k) vertices and (formula presented) edges. Finally, we study Bounded Cactus Contraction (Bounded CC) which takes as input a graph G and integers k and. The goal is to decide whether there exists a subset (formula presented) of size at most k such that G/F is a cactus graph with at most (formula presented) leaf blocks in the corresponding block decomposi-tion. For Bounded CC we design a kernel with (formula presented) vertices and (formula presented) edges. We complement our results by giving kernelization lower bounds for Bounded TC, Bounded OTC and Bounded CC by showing that unless (formula presented) the size of the kernel we obtain is optimal. © Springer International Publishing AG 2017.

Paths to trees and cacti

Commonly used encryption methods treat the plaintext merely as a stream of bits, disregarding any specific format that the data might have. In many situations, it is desirable and essential to have the ciphertext follow the same format as the plaintext. Moreover, ciphertext length expansion is also not allowed in these situations. Encryption of credit card numbers and social security numbers are the two most common examples of this requirement. Format-Preserving Encryption (FPE) is a symmetric key cryptographic primitive that is used to achieve this functionality. Initiated by the work of Black and Rogaway (CT-RSA 2002), many academic solutions have been proposed in literature that have focused on designing efficient FPE schemes. However, almost all the existing FPE schemes are based on Feistel construction and have efficiency issues. In this work, we propose a new family of efficient FPE schemes that are Substitution-Permutation (SP) based constructions at their core. We term it as SPF family of FPE schemes. All the underlying SP transformations in these constructions have been defined such that they preserve the format of the data. We then demonstrate an instance of our construction applicable for digits. We show that our scheme is at least 5 times more efficient than existing FPE designs for most of the practical applications. © Springer International Publishing AG 2017.

SPF: A new family of efficient format-preserving encryption algorithms

Zone extraction is acclaimed as a significant pre-processing step in handwriting analysis. This paper presents a new method for separating ascenders and descenders from an unconstrained handwritten word and identifying its core-region. The method estimates correct core-region for complexities like long horizontal strokes, skewed words, first letter capital, hill and dale writing, jumping baselines and words with long descender curves, cursive handwriting, calligraphic words, title case words, very short words as shown in Fig. 1. It extracts two envelops from the word image and selects sample points that constitute the core region envelop. The method is tested on CVL, ICDAR-2013, ICFHR-2012, and IAM benchmark datasets of handwritten words written by multiple writers. We also created our own dataset of 100 words authored by 2 writers comprising all the above mentioned handwriting complexities. Due to non-availability of the Ground Truth for core-region extraction we created it manually for all the datasets. Our work reports an accuracy of 90.16% for correctly identifying all the three zones on 17,100 Latin words written by 802 individuals. Promising results are obtained by our core-region detection method when compared with the current state of the art methods. © 2017 IEEE.

Proceedings of the International Conference on Document Analysis and Recognition, ICDAR

Core Region Detection for Off-Line Unconstrained Handwritten Latin Words Using Word Envelops

Segmentation is often performed on medical images for identifying diseases in clinical evaluation. Hence it has become one of the major research areas. Conventional image segmentation techniques are unable to provide satisfactory segmentation results for medical images as they contain irregularities. They need to be pre-processed before segmentation. In order to obtain the most suitable method for medical image segmentation, we propose MIST (Medical Image Segmentation Tool), a two stage algorithm. The first stage automatically generates a binary marker image of the region of interest using mathematical morphology. This marker serves as the mask image for the second stage which uses GrabCut to yield an efficient segmented result. The obtained result can be further refined by user interaction, which can be done using the proposed Graphical User Interface (GUI). Experimental results show that the proposed method is accurate and provides satisfactory segmentation results with minimum user interaction on medical as well as natural images. © 2017 Elsevier Ltd

Fulltext

Computers in Biology and Medicine

An interactive medical image segmentation framework using iterative refinement

In the highly growing digital world, information sharing is important common. But it must be secured else adversary can either intercept or delete partially/fully information. Hence, it will be useless to specific human but it may be helpful to adversary. In today's road market, vehicles are most useful to assist persons through broadcasting vehicles regarding information such that current position, traffic, environment, etc. to other vehicles in a secure vehicular communication. We will use public-private key mechanism to achieve this scenario in a public domain. The main goal of this paper is to improve security of exchanging message in terms of confidentiality, authentication, access & availability. Hence, we have suggested an approach to fulfill these security needs for information transmission between two or more vehicles in doubtful network infrastructure. Thus, vehicles can send/receive their important credentials through our secure message exchanging algorithm. Because, in our algorithm, we have used idea of public-private key encryption-decryption mechanism in the computation of messages which is sent over public channel. © 2016 IEEE.

IEEE Region 10 Annual International Conference, Proceedings/TENCON

Secure message transmission algorithm for Vehicle to Vehicle (V2V) communication

The personal photo collections are becoming significant in our day today existence. The challenge is to precisely intuit user’s complex and transient interests and to accordingly develop an adaptive and automated personalized photo management system which efficiently manages and organizes personal photos. This is increasingly gaining importance as it will be required to browse, search and retrieve efficiently the relevant information frompersonal collections whichmay extend from many years. Significance and relevance for the user also may undergo temporal and crucial shifts which need to be continually logged to generate patterns. The cloud paradigm makes available the basic platform but a system needs to be built wherein a personalized service with ability to capture diversity is guaranteed even when the training data size is small. An Artificial Immune Hybrid Photo Album Classifier (AIHPAC) is proposed using the nonlinear biological properties of Human Immune Systems. The system does event based clustering for an individual with embedded feature selection. The model is self learning and self evolving. The efficacy of the proposed method is efficiently demonstrated by the experimental results. © Springer Science+Business Media Singapore 2017.