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

Proportions of different phases (phase fraction) in the microstructures determine the quality of dual phase (DP) steel. So, calculation of phase fraction in the microstructures of steel samples is important for quality assurance. Manual calculation of phase fraction involves Le Pera etching of steel which is time consuming and dependent on operator efficiency. Calculation of phase fraction from Le Pera etched samples requires cumbersome manual observations. Nital etching is a faster alternative to Le Pera etching. However, due to lack of visually discriminative information, different phases cannot be identified manually from nital images. We propose a novel method for automatic calculation of phase fractions in steel microstructures from nital images using machine learning techniques. We show that regional contour patterns and local entropy (which cannot be evaluated manually) of regions of nital images are related to the formation process of the phases. We design a method that automatically evaluates regional contour patterns and local entropy from nital images of DP steel. Subsequently, we construct a random forest classifier that uses regional contour patterns and local entropy as features for classification of different phases. Our method is ~150 times faster than manual classification. Experiments show close to 90% accuracy in classification. © The Institution of Engineering and Technology 2018.

World University root node

IET Image Processing

Calculation of phase fraction in steel microstructure images using random forest classifier

Road safety is a critical issue the world-over, and the problem is particularly acute in developing countries, where the combination of crowding, inadequate roads, and driver indiscipline serves up a deadly cocktail. We believe that mobile devices can play a positive role in this context by detecting dangerous conditions and providing feedback to enable timely redressal of potential dangers. This paper focuses on a specific problem that is responsible for many accidents in developing countries: the stopping behaviour of buses especially in the vicinity of bus stops. For instance, buses could arrive at a bus stop but continue rolling forward instead of coming to a complete halt, or could stop some distance away from the bus stop, possibly even in the middle of a busy road. Each of these behaviours can result in injury or worse to people waiting at a bus stop as well as to passengers boarding or alighting from buses. We present FullStop, a smartphone-based system to detect safety risks arising from bus stopping behaviour, as described above. We show that the GPS and inertial sensors are unable to perform the fine-grained detection needed, by themselves. Therefore, FullStop is based on the view obtained from looking out to the front of the vehicle using the camera of a smartphone that is mounted on the front windshield. Using optical flow vectors, with several refinements, FullStop running on a smartphone is able to effectively detect safety-related situations such as a rolling stop or stopping at a location that is displaced laterally relative to the designated bus stop. © 2018 IEEE.

2018 10th International Conference on Communication Systems and Networks, COMSNETS 2018

FullStop: Tracking unsafe stopping behaviour of buses

Fourier Transform Holography (FTH) encodes object information as the Fourier spectrum in an interference pattern with the help of a reference point source. Hence, a single Fourier transform (FT) operation is capable to provide a complete object from a digitally reconstructed FTH. However, high intensity values in the central region of the FTH demands central blocking due to the limited dynamic range of a detector in a variety of applications. The central blocking causes the loss of frequency content and hence affects the reconstruction result. Still, it can be properly recovered using signal processing techniques such as Compressive Sensing (CS) or Sparse signal processing. This paper investigates the role of CS for centrally blocked FTH using various kind of beam stops and provide a detailed comparison between usual FT reconstruction and technique equipped with CS. We also provide a detail procedure to apply the CS reconstruction in the FTH and compared the results with and without CS. It is established that an object can be recovered from limited size of a recorded hologram with the constraint that object has to be sparse in known basis. Quantitative analysis has been carried out to compare the quality of reconstruction using inverse Fourier transform (IFT) and CS techniques. Quality of reconstruction is assessed by evaluating measures such as the overall visibility, efficiency and PSNR of reconstruction. The finding highlights that the CS reconstruction is better than that of IFT reconstruction. Both simulation and experimental results demonstrate the effectiveness and robustness of the CS-based approach. © 2018 Elsevier Ltd

Optics and Lasers in Engineering

Reconstruction of apertured Fourier Transform Hologram using compressed sensing

Digital retouching of face images is becoming more widespread due to the introduction of software packages that automate the task. Several researchers have introduced algorithms to detect whether a face image is original or retouched. However, previous work on this topic has not considered whether or how accuracy of retouching detection varies with the demography of face images. In this paper, we introduce a new Multi-Demographic Retouched Faces (MDRF) dataset, which contains images belonging to two genders, male and female, and three ethnicities, Indian, Chinese, and Caucasian. Further, retouched images are created using two different retouching software packages. The second major contribution of this research is a novel semi-supervised autoencoder incorporating 'sub-class' information to improve classification. The proposed approach outperforms existing state-of-the-art detection algorithms for the task of generalized retouching detection. Experiments conducted with multiple combinations of ethnicities show that accuracy of retouching detection can vary greatly based on the demographics of the training and testing images. © 2017 IEEE.

IEEE International Joint Conference on Biometrics, IJCB 2017

Demography-based facial retouching detection using subclass supervised sparse autoencoder

Autoencoders are unsupervised deep learning models used for learning representations. In literature, autoencoders have shown to perform well on a variety of tasks spread across multiple domains, thereby establishing widespread applicability. Typically, an autoencoder is trained to generate a model that minimizes the reconstruction error between the input and the reconstructed output, computed in terms of the Euclidean distance. While this can be useful for applications related to unsupervised reconstruction, it may not be optimal for classification. In this paper, we propose a novel Supervised COSMOS Autoencoder which utilizes a multi-objective loss function to learn representations that simultaneously encode the (i) "similarity" between the input and reconstructed vectors in terms of their direction, (ii) "distribution" of pixel values of the reconstruction with respect to the input sample, while also incorporating (iii) "discriminability" in the feature learning pipeline. The proposed autoencoder model incorporates a Cosine similarity and Mahalanobis distance based loss function, along with supervision via Mutual Information based loss. Detailed analysis of each component of the proposed model motivates its applicability for feature learning in different classification tasks. The efficacy of Supervised COSMOS autoencoder is demonstrated via extensive experimental evaluations on different image datasets. The proposed model outperforms existing algorithms on MNIST, CIFAR-10 and SVHN databases. It also yields state-of-the-art results on CelebA, LFWA, Adience and IJB-A databases for attribute prediction and face recognition, respectively.

arXiv preprint arXiv:1810.

Supervised COSMOS Autoencoder: Learning Beyond the Euclidean Loss!

In this paper, we improve the performance of the recently proposed Direct Query Classifier (DQC). The (DQC) is a classifier based retrieval method and in general, such methods have been shown to be superior to the OCR-based solutions for performing retrieval in many practical document image datasets. In (DQC), the classifiers are trained for a set of frequent queries and seamlessly extended for the rare and arbitrary queries. This extends the classifier based retrieval paradigm to an unlimited number of classes (words) present in a language. The (DQC) requires indexing cut-portions (n-grams) of the word image and DTW distance has been used for indexing. However, DTW is computationally slow and therefore limits the performance of the (DQC). We introduce query specific DTW distance, which enables effective computation of global principal alignments for novel queries. Since the proposed query specific DTW distance is a linear approximation of the DTW distance, it enhances the performance of the (DQC). Unlike previous approaches, the proposed query specific DTW distance uses both the class mean vectors and the query information for computing the global principal alignments for the query. Since the proposed method computes the global principal alignments using n-grams, it works well for both frequent and rare queries. We also use query expansion (QE) to further improve the performance of our query specific DTW. This also allows us to seamlessly adapt our solution to new fonts, styles and collections. We have demonstrated the utility of the proposed technique over 3 different datasets. The proposed query specific DTW performs well compared to the previous DTW approximations. © 2018 by the authors. Licensee MDPI, Basel, Switzerland.

Fulltext

Journal of Imaging

Efficient query specific dtw distance for document retrieval with unlimited vocabulary

More study is needed to make VANETs more relevant. Opportunistic routing (OR) is a new model that has been proposed for wireless networks. OR has emerged from the research communities because of its ability to increase the performance of wireless networks. It benefits from the broadcast characteristic of wireless mediums to improve network performance. The basic function of OR is its ability to overhear the transmitted packet and to coordinate among relaying nodes. In this chapter, an exhaustive survey of existing OR protocols is done by considering various factors. More precisely, existing secure OR protocols are deliberated. Future directions of research are also included, which provide a superior way to overcome some of the limitations of these existing protocols. Through this detailed survey, an outline and in-depth knowledge of existing OR protocols can be acquired. © 2018, IGI Global.

Handbook of Research on Pattern Engineering System Development for Big Data Analytics

Secure opportunistic routing for vehicular adhoc networks

We formulate a multi-directional broadcast (MDB) storm problem arising in dense vehicular ad-hoc networks (VANETs) when the multiple nodes (moving in multiple directions) forward broadcast packets meet at the road intersections, resulting in severe packet collisions inducing delays at medium access control. In this work, we have proposed a mechanism for partitioning the graph of vehicles in an intersection into multiple bipartite directional sub-graphs, such that each sub-graph aggregates messages using short range communication and make one long range communication of aggregated message. The k balanced graph partitioning problem contains partitions of size ≤ |V|/k nodes. For a graph G = (V, E), a partitioning P, is (k, 1 + ε) balanced if V is partitioned into k disjoint subsets each containing at most (1 + ε)n/k vertices. Our proposed approximation algorithm for intersection-cast problem uses a balanced partition with Θ (log^2 n) approximation for balance constant, v > 1. We have given simulation results for the performance analysis of our intersection-cast protocol compared to the existing competitive schemes and found improvement in terms of broadcast success rate, reachability and message overhead in the networks. Copyright © 2018 Inderscience Enterprises Ltd.

International Journal of Communication Networks and Distributed Systems

IntersectionCast: Approximation algorithm for multi-directional broadcast storm in VANETs

With the exponential increase in number of connected sensor devices for various smart home applications, a huge amount of heterogeneous data are generated every day. Collection technique plays a crucial role in defining the application QoS. In this paper, we propose a novel data collection framework to serve different applications used for critical infrastructure monitoring. The proposed framework adopts a context-aware forwarding strategy to ensure higher delivery reliability. The overall cost of data collection is significantly low due to the incorporation of in-network data aggregation and fusion on demand basis. The effectiveness of the proposed scheme has also been established using simulation results. © 2018 Association for Computing Machinery.

Proceedings of the Annual International Conference on Mobile Computing and Networking, MOBICOM

Context aware data collection framework for critical infrastructure monitoring system

Cloud computing is a developing technology that is yet unclear to many security issues. Data in the untrusted clouds can be encrypted using encryption algorithm. Randomizing this data provides more security which can be achieved by padding concept in the cloud. In this paper, the user's data is encrypted using padding scheme, called Optimal Asymmetric Encryption Padding (OAEP) together with Hybrid Encryption algorithm that is based on RSA (i.e., HE-RSA), in order to allow multiple parties to compute a function on their inputs while preserving Integrity and Confidentiality. The Homomorphic Encryption(HE) is performed on the encrypted data without decrypting it in computationally powerful clouds and the Secure Multi-Party Computation (SMPC) can be used in the cloud to ensure security and privacy of the users. In this paper, we have proposed a scheme that integrates the multi-party computation with homomorphic encryption to allow calculations of encrypted data without decryption. The cryptographic techniques used in our cloud model are described and the overheads are compared with Homomorphic Encryption and Multi-Party Computation. © 2018 IEEE.