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

Watermarking plays a very important role in providing authentication, ownership and transmission of secret information. The existing techniques of watermarking in literature are based on either spatial domain or transformation domain. Human brain consists of large number of neurons which are capable of doing paralleling tasking accurately. This resulted in the evaluation of neural network architectures, providing wide range of well connected features representing the input of the network. Convolutional Neural Networks(CNN) which were evolved in early 90s became popular and are being in use in wide range of tasks like classification, detection, recognition, patch matching. In this paper, we propose a digital image watermarking technique using auto-encoder based CNN which is robust to different noises and attacks like salt & pepper, Gaussian and JPEG effect. The proposed method, to the best of our knowledge, is the very first attempt of CNN in the domain of watermarking. We compare the performance of the proposed technique with the existing techniques which are based on spatial domain and transform domain. We show that the proposed method of watermarking using auto encoder based CNN (ACNNWM) gives better or on par results with the existing methods. © 2015 IEEE.

World University root node

2015 IEEE Workshop on Computational Intelligence: Theories, Applications and Future Directions, WCI 2015

A robust digital image watermarking technique using auto encoder based convolutional neural networks

Automated detection of visually salient regions is an active area of research in computer vision. Salient regions can serve as inputs for object detectors as well as inputs for region-based registration algorithms. In this paper, we consider the problem of speeding up computationally intensive bottom-up salient region detection in 3D medical volumes. The method uses the Kadir-Brady formulation of saliency. We show that in the vicinity of a salient region, entropy is amonotonically increasing function of the degree of overlap of a candidate window with the salient region. This allows us to initialize a sparse seed point grid as the set of tentative salient region centers and iteratively converge to the local entropy maxima, thereby reducing the computation complexity compared to the Kadir-Brady approach of performing this computation at every point in the image. We propose two different approaches for achieving this. The first approach involves evaluating entropy in the four quadrants around the seed point and iteratively moving in the direction that increases entropy. The second approach we propose makes use of mean shift tracking framework to affect entropy maximizing moves. Specifically, we propose the use of uniform pmf as the target distribution to seek high entropy regions.We demonstrate the use of our algorithm on medical volumes for left ventricle detection in PET images and tumor localization in brain MR sequences.

Machine Intelligence and Signal Processing

In this work we present a framework for recognizing writer for a handwritten word. We make use of allographic features at sub-word level. Our work is motivated by previous techniques which make use of a codebook. However, instead of encoding the features using the code-words, we exploit the discriminative properties of features that belong to the same cluster, in a supervised approach. We are able to achieve writer identification rates close to 63% on the handwritten words drawn from a dataset by 10 writers. Our work has application in scenarios where multiple writers write/annotate on the same page. © Springer International Publishing AG 2017.

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

Writer identification for handwritten words

People with speech disabilities communicate in sign language and therefore have trouble in mingling with the able-bodied. There is a need for an interpretation system which could act as a bridge between them and those who do not know their sign language. A functional unobtrusive Indian sign language recognition system was implemented and tested on real world data. A vocabulary of 140 symbols was collected using 18 subjects, totalling 5041 images. The vocabulary consisted mostly of two-handed signs which were drawn from a wide repertoire of words of technical and daily-use origins. The system was implemented using Microsoft Kinect which enables surrounding light conditions and object colour to have negligible effect on the efficiency of the system. The system proposes a method for a novel, low-cost and easy-to-use application, for Indian Sign Language recognition, using the Microsoft Kinect camera. In the fingerspelling category of our dataset, we achieved above 90% recognition rates for 13 signs and 100% recognition for 3 signs with overall 16 distinct alphabets (A, B, D, E, F, G, H, K, P, R, T, U, W, X, Y, Z) recognised with an average accuracy rate of 90.68%. © 2016, Indian Academy of Sciences.

Sadhana - Academy Proceedings in Engineering Sciences

Nearest neighbour classification of Indian sign language gestures using kinect camera

The task of automatic image annotation involves assigning relevant multiple labels/tags to query images based on their visual content. One of the key challenge in multilabel image annotation task is the class imbalance problem where frequently occurring labels suppress the participation of rarely occurring labels. In this paper, we propose to exploit the multi-scale behavior in hypergraph heat diffusion framework for the automatic image annotation task. The proposed novel technique enables to model the higher order relationship among images in the feature space and provides a multi-scale label diffusion mechanism to address the class imbalance problem in the data. © 2016 ACM.

ICMR 2016 - Proceedings of the 2016 ACM International Conference on Multimedia Retrieval

Image annotation using multi-scale hypergraph heat diffusion framework

This paper deals with several new methodologies and concepts in the area of rough set theoretic granular computing which are then applied in video tracking. A new concept of neighborhood granule formation over images is introduced here. These granules are of arbitrary shapes and sizes unlike other existing granulation techniques and hence more natural. The concept of rough-rule base is used for video tracking to deal with the uncertainties and incompleteness as well as to gain in computation time. A new neighborhood granular rough rule base is formulated which proves to be effective in reducing the indiscernibility of the rule-base. This new rule-base provides more accurate results in the task of tracking. Two indices to evaluate the performance of tracking are defined. These indices do not need ground truth information or any estimation technique like the other existing ones. All these features are demonstrated with suitable experimental results. © 2015, Springer Science+Business Media Dordrecht.

Natural Computing

Neighborhood granules and rough rule-base in tracking

There are monocular depth cues present in images or videos that aid in depth perception in two-dimensional images or videos. Our objective is to preserve the defocus depth cue present in the videos along with the salient regions during compression application. A method is provided for opportunistic bit allocation during the video compression using visual saliency information comprising both the image features, such as color and contrast, and the defocus-based depth cue. The method is divided into two steps: saliency computation followed by compression. A nonlinear method is used to combine pure and defocus saliency maps to form the final saliency map. Then quantization values are assigned on the basis of these saliency values over a frame. The experimental results show that the proposed scheme yields good results over standard H.264 compression as well as pure and defocus saliency methods. © 2016 SPIE and IS&T.

Journal of Electronic Imaging

Defocus cue and saliency preserving video compression

With the rapid increase of multimedia data, textual content in an image has become a very important source of information for several applications like navigation, image search and retrieval, image understanding, captioning, machine translation and several others. Scene text localization is the first step towards such applications and most current methods focus on generating a small set of high precision detectors rather than obtaining large set of detections covering all text patches. In this work we propose a novel hybrid framework for text localization which uses character level recognition recursively in a feedback mechanism to refine text patches and reduce false positives. We use popular MSER algorithm at multiple scales as an initial region proposal algorithm and several filtering stages recursively to improve precision as well as maximize recall. We aim at achieving high recall rather than achieving higher precision since several robust word recognition systems are already available. The word recognition systems are mature enough to produce highly accurate results if provided with maximum amount of regions rather than providing small set of highly precise text patches and losing several other text regions. The main contribution of this paper is the use of character recognizer within a novel feedback mechanism to recursively search for text regions in the neighborhood of previously detected text patches. Using 3 publicly available benchmark datasets (ICDAR2011, MSRA TD-500 and OSTD), we demonstrate the efficacy of the proposed framework for text localization. © 2016 IEEE.

Proceedings - International Conference on Pattern Recognition

Recognition based text localization from natural scene images

Marketers use segmentation as an important tool to better understand and effectively target customers by adopting marketing strategies catering to the needs and characteristics of each segment. Traditionally, customer segmentation using clustering algorithms is performed using a single attribute set. However, there may be multiple meaningful natural customer groupings possible if independent subsets of attributes were considered while discovering the clusters. Multi-view clustering provides a meaningful way to combine groupings in different feature subsets by assigning customers to multiple segments, representing different perspectives of customer behavior. In this paper we propose a novel method for performing multi-view clustering wherein multiple groupings are generated using a non-parametric clustering algorithm and are then combined and visualized using a cross-tab/sunburst based visualization scheme. We also demonstrate the effectiveness of the proposed approach by applying it to a variety of real-world problems related to mobile subscriber segmentation. © 2016 IEEE.

Proceedings - CBI 2016: 18th IEEE Conference on Business Informatics

A Multi-view Non-parametric Clustering Approach to Mobile Subscriber Segmentation

This paper presents a camera model to deal with underwater scene reconstruction from multiple images. Effects due to change in path of light ray at the medium interface is modelled for a general medium with unknown refractive index. Our model calculates the refractive index of the medium and simultaneously removes the geometric refraction effects in images using point correspondences in image pairs. With known internal parameters of the camera we find the external parameters of the camera and then 3-D reconstruction is obtained. © 2015 IEEE.