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

Face recognition algorithms generally use 2D images for feature extraction and matching. In order to achieve better performance, 3D faces captured via specialized acquisition methods have been used to develop improved algorithms. While such 3D images remain difficult to obtain due to several issues such as cost and accessibility, RGB-D images captured by low cost sensors (e.g. Kinect) are comparatively easier to acquire. This research introduces a novel face recognition algorithm for RGB-D images. The proposed algorithm computes a descriptor based on the entropy of RGB-D faces along with the saliency feature obtained from a 2D face. The probe RGB-D descriptor is used as input to a random decision forest classifier to establish the identity. This research also presents a novel RGB-D face database pertaining to 106 individuals. The experimental results indicate that the RGB-D information obtained by Kinect can be used to achieve improved face recognition performance compared to existing 2D and 3D approaches. © 2013 IEEE.

World University root node

IEEE 6th International Conference on Biometrics: Theory, Applications and Systems, BTAS 2013

On RGB-D face recognition using Kinect

Face recognition from still face images suffers due to intrapersonal variations caused by pose, illumination, and expression that degrade the performance. On the other hand, videos provide abundant information that can be leveraged to compensate the limitations of still face images and enhance face recognition performance. This paper presents a video based face recognition algorithm that computes a discriminative video signature as an ordered list of still face images. The video signature embeds diverse intra-personal and temporal variations across multiple frames, thus facilitates matching two videos with large variations. Two videos are matched by comparing their discriminative signatures using the Kendall tau similarity distance measure. Performance comparison with the benchmark results and a commercial face recognition system on the publicly available YouTube faces database show the efficacy of the proposed video based face recognition algorithm. © 2013 IEEE.

2013 IEEE International Conference on Image Processing, ICIP 2013 - Proceedings

On rank aggregation for face recognition from videos

Proceedings - 2013 International Conference on Biometrics, ICB 2013

Program chairs' welcome to ICB 2013

This paper presents the design of STAR (Spatio-Temporal Analysis and Retrieval), an unsupervised Content Based Video Retrieval (CBVR) System. STAR's key insight and primary contribution is that it models video content using a joint spatio-temporal feature representation and retrieves videos from the database which have similar moving object and trajectories of motion. Foreground moving blobs from a moving camera video shot are extracted, along with a trajectory for camera motion compensation, to form the space-time volume (STV). The STV is processed to obtain the EMST-CSS representation, which can discriminate across different categories of videos. Performance of STAR has been evaluated qualitatively and quantitatively using precision-recall metric on benchmark video datasets having unconstrained video shots, to exhibit efficiency of STAR. © 2013 IEEE.

2013 4th National Conference on Computer Vision, Pattern Recognition, Image Processing and Graphics, NCVPRIPG 2013

STAR: A Content Based Video Retrieval system for moving camera video shots

Compressed Sensing (CS) is a new signal acquisition technology that enables the reduction in the number of measurements required for recovery of sparse or compressible signal in some basis. Signal recovery is carried out with the help of optimization techniques. Here we are conducting a review of some of these algorithms and trying to find the best algorithm to make use of that in launch vehicle telemetry system. © 2013 IEEE.

2013 Annual International Conference on Emerging Research Areas, AICERA 2013 and 2013 International Conference on Microelectronics, Communications and Renewable Energy, ICMiCR 2013 - Proceedings

Feasibility study of applying compressed sensing recovery algorithms for launch vehicle telemetry

Face verification, though for humans seems to be an easy task, is a long-standing research area. With challenging covariates such as disguise or face obfuscation, automatically verifying the identity of a person is assumed to be very hard. This paper explores the feasibility of face verification under disguise variations using multi-spectrum (visible and thermal) face images. We propose a framework, termed as Aravrta1, which classifies the local facial regions of both visible and thermal face images into biometric (regions without disguise) and non-biometric (regions with disguise) classes. The biometric patches are then used for facial feature extraction and matching. The performance of the algorithm is evaluated on the IHTD In and Beyond Visible Spectrum Disguise database that is prepared by the authors and contains images pertaining to 75 subjects with different kinds of disguise variations. The experimental results suggest that the proposed framework improves the performance compared to existing algorithms, however there is a need for more research to address this important covariate. © 2013 IEEE.

Disguise detection and face recognition in visible and thermal spectrums

A covariate in face recognition can be defined as an effect that independently\r\nincreases the intra-class variability or decreases the inter-class variability or\r\nboth. Covariates such as pose, illumination, expression, aging and disguise\r\nare established and extensively studied in literature and are categorized as\r\nexisting covariates of face recognition. However, ever increasing applications\r\nof face recognition have instigated many new and exciting scenarios such as\r\nmatching forensic sketches to mug-shot photos, faces altered due to plastic\r\nsurgery, low resolution surveillance images and individual from videos. These\r\ncovariates are categorized as emerging covariates of face recognition, which\r\nis the primary emphasis of this dissertation. One of the important cues in\r\nsolving crimes and apprehending criminals is matching forensic sketches with\r\ndigital face images. The first contribution of this dissertation is a memetically\r\noptimized multi-scale circular Weber's local descriptor (MCWLD) for matching\r\nforensic sketches with digital face images. This dissertation presents an\r\nautomated algorithm to extract discriminative information from local regions\r\nof both sketches and digital images using MCWLD. An evolutionary memetic\r\noptimization is proposed to assign optimal weights to every local facial region\r\nto boost the identification performance. Since, forensic sketches and digital images\r\ncan be of poor quality, a pre-processing technique is also used to enhance\r\nthe quality of images. Results on different sketch databases, including forensic\r\nsketch database, illustrate the efficacy of the proposed algorithm. Widespread\r\nacceptability and use of biometrics for person authentication has instigated\r\nseveral techniques for evading identification such as altering facial appearance\r\nusing surgical procedures. These procedures modify both the shape and texture\r\nof facial features to varying degrees and thus degrade the performance\r\nof face recognition when matching pre- and post-surgery images. The second\r\ncontribution of this dissertation is a multi-objective evolutionary granular algorithm\r\nfor matching face images altered due to plastic surgery procedures.\r\nThe algorithm first generates non-disjoint face granules at multiple levels of\r\ngranularity. The granular information is assimilated using a multi-objective genetic\r\nalgorithm that simultaneously optimizes the selection of feature extractor\r\nfor each face granule along with the weights of individual granules. On IIIT-D\r\nplastic surgery database, the proposed algorithm yields the state-of-the-art performance.\r\nFace recognition performance degrades when a low resolution face\r\nimage captured in unconstrained settings, such as surveillance, is matched with\r\nhigh resolution gallery images. The primary challenge is to extract discriminative\r\nfeatures from the limited biometric content in low resolution images\r\nand match it with information-rich high resolution face images. The problem\r\nof cross-resolution face matching is further alleviated when there is limited\r\nlabeled low resolution training data. The third contribution of this dissertation\r\nis co-transfer learning framework, a cross pollination of transfer learning\r\nand co-training paradigms, for enhancing the performance of cross-resolution\r\nface recognition. The transfer learning component transfers the knowledge\r\nthat is learned while matching high resolution face images during training\r\nfor matching low resolution probe images with high resolution gallery during\r\ntesting. On the other hand, co-training component facilitates this knowledge\r\ntransfer by assigning pseudo labels to unlabeled probe instances in the target\r\ndomain. Experiments on a synthetic, three low resolution surveillance\r\nquality face databases and real world examples show the efficacy of the proposed\r\nco-transfer learning algorithm as compared to other approaches. Due\r\nto prevalent applications and availability of large intra-personal variations,\r\nvideos have gained significant attention for face recognition. Unlike still face\r\nimages, videos provide abundant information that can be leveraged to compensate\r\nfor variations in intra-personal variations and enhance face recognition\r\nperformance. The fourth contribution of this dissertation is a video based face\r\nrecognition algorithm which computes a discriminative video signature as an\r\nordered (ranked) list of still face images from a large dictionary. A three stage\r\napproach is developed for optimizing ranked lists across multiple video frames\r\nand fusing them into a single composite ordered list to compute the video signature.\r\nThe signature embeds diverse intra-personal variations and facilitates in\r\nmatching two videos across large variations. Results obtained on Youtube and\r\nMBGC v2 video databases show the effectiveness of the proposed algorithm.

Emerging Covariates of Face Recognition

Int J Curr Microbiol Appl Sci

Original Research Article Comparative study of ESBL producing Escherichia coli in OPD and IPD patients of urinary tract infections

Message from the program chairs

Geographic Perimeter Stateless Routing (GPSR) makes use of location information to forward packets greedily. Nodes need to keep only this local information, hence called as stateless. There are two graphs used for the planarization namely the Gabriel Graph (GG) and the Relative Neighbourhood Graph (RNG) but they were very expensive and had to be run for each node whenever greedy forwarding failed. So it was further replaced by CLDP, however it also suffered from drawbacks both in terms of cost and complications. Here, we propose a cost-effective scheme which makes use of k-hop neighbourhood knowledge and this heuristic is applied when greedy forwarding fails and is used instead of face routing and planarization. Moreover it proves to be better than planarization in most of the cases. © 2013 IEEE.