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

Due to the strong capacity of deep learning in handling unstructured data, it has been utilized for the task of single image super-resolution (SISR). These algorithms have shown promising results for small scale super-resolution but are not robust to large scale super-resolution. In addition, these algorithms are computationally complex and require high-end computational devices. Developing large-scale super-resolution framework finds its application in smart-phones as these devices have limited computational power. In this context, we present a novel light-weight architecture-Gradually growing Residual and self-Attention based Dense Deep Back Projection Network (GRAD-DBPN) for large scale image super-resolution (SR). The network is made of cascaded self-Attention based Residual Dense Deep Back Projection Network (ARD-DBPN) blocks to perform super-resolution gradually. Where each block performs 2X super-resolution and fine tuned in an end to end manner. The residual architecture facilitates the faster convergence of network and overcomes the issue of vanishing gradient. Experimental results on different benchmark data-set have been presented to compare the efficacy and effectiveness of the architecture. © 2019, Springer Nature Switzerland AG.

World University root node

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

Gradually Growing Residual and Self-attention Based Dense Deep Back Projection Network for Large Scale Super-Resolution of Image

This paper is concerned with the development of techniques for the recognition of ornamental characters motivated by the perceptual processes involved in humans. To understand the perceptual process, we have performed the eye-tracking experiment to recognize the special set of characters, with artistic variations in character structure and form. The novelty of this paper is the use of human visual fixations to supervise the intermediate layers of the convolutional neural network. From the results obtained, we found that the network performs better when trained with the assistance of fixation information compared to the network trained without eye fixations.

2019 International Conference on Document Analysis and Recognition Workshops, ICDARW 2019

Intelligent Identification of Ornamental Devanagari Characters Inspired by Visual Fixations

Objective: Segmentation of anatomical structures in ultrasound images requires vast radiological knowledge and experience. Moreover, the manual segmentation often results in subjective variations, therefore, an automatic segmentation is desirable. We aim to develop a fully convolutional neural network (FCNN) with attentional deep supervision for the automatic and accurate segmentation of the ultrasound images. Method: FCNN/CNNs are used to infer high-level context using low-level image features. In this paper, a sub-problem specific deep supervision of the FCNN is performed. The attention of fine resolution layers is steered to learn object boundary definitions using auxiliary losses, whereas coarse resolution layers are trained to discriminate object regions from the background. Furthermore, a customized scheme for downweighting the auxiliary losses and a trainable fusion layer are introduced. This produces an accurate segmentation and helps in dealing with the broken boundaries, usually found in the ultrasound images. Results: The proposed network is first tested for blood vessel segmentation in liver images. It results in F1 score, mean intersection over union, and dice index of 0.83, 0.83, and 0.79, respectively. The best values observed among the existing approaches are produced by U-net as 0.74, 0.81, and 0.75, respectively. The proposed network also results in dice index value of 0.91 in the lumen segmentation experiments on MICCAI 2011 IVUS challenge dataset, which is near to the provided reference value of 0.93. Furthermore, the improvements similar to vessel segmentation experiments are also observed in the experiment performed to segment lesions. Conclusion: Deep supervision of the network based on the input-output characteristics of the layers results in improvement in overall segmentation accuracy. Significance: Sub-problem specific deep supervision for ultrasound image segmentation is the main contribution of this paper. Currently the network is trained and tested for fixed size inputs. It requires image resizing and limits the performance in small size images. © 2019 IEEE.

IEEE Transactions on Biomedical Engineering

Ultrasound Image Segmentation: A Deeply Supervised Network With Attention to Boundaries

Recent advancements in deep learning have significantly increased the capabilities of face recognition. However, face recognition in an unconstrained environment is still an active research challenge. Covariates such as pose and low resolution have received significant attention, but 'disguise' is considered an onerous covariate of face recognition. One primary reason for this is the unavailability of large and representative databases. To address the problem of recognizing disguised faces, we propose an active learning framework A-LINK*, that intelligently selects training samples from the target domain data, such that the decision boundary does not overfit to a particular set of variations, and better generalizes to encode variability. The framework further applies domain adaptation with the actively selected training samples to fine-tune the network. We demonstrate the effectiveness of the proposed framework on DFW and Multi-PIE datasets with state-of-the-art models such as LCSSE and DenseNet. © 2019 IEEE.

2019 IEEE 10th International Conference on Biometrics Theory, Applications and Systems, BTAS 2019

A-LINK: Recognizing Disguised Faces via Active Learning based Inter-Domain Knowledge

With the advent of deep learning architectures, the performance of face recognition has witnessed significant improvements. However, this has also necessitated the requirement of large labeled training database. While approaches exist to utilize labeled or unlabeled data from related domains, in this paper, we present a collaborative learning framework that utilizes the availability of both labeled and unlabeled data along with the presence of multiple experts, to improve the performance of face analysis related tasks. The proposed Label Consistent Deep Collaborative Learning (LC-DECAL) framework makes use of label consistency, transfer learning, ensemble learning, and co-training for training a deep neural network for the target domain. The efficacy of the proposed algorithm is demonstrated with two existing Convolutional Neural Network architectures, DenseNet and ResNet, via experiments on multiple face databases, namely YTF, PaSC Handheld, PaSC Control, CelebA, and LFW-a. Experimental results show that the proposed framework yields comparable results to state-of-the-art results on all the databases. © 2019 IEEE.

LC-DECAL: Label Consistent Deep Collaborative Learning for Face Recognition

In complex systems like'Internet of Things' and'Cyber Physical Systems', reliable communication is made possible by the use of error control codes like turbo codes. Such systems are vulnerable to attack by Trojans during the design or fabrication phases of IC manufacturing. This leads to a severe performance degradation of the entire system. This paper aims at analyzing the effect of a sign-bit-flipping Trojan on a turbo coded communication system. Simulations were carried out on a conventional iterative turbo decoder and the results point out that the Trojan affects the performance of the decoder at SNRs greater than -2 dB. This performance degradation is seen even with minimal activation of the Trojan and a finite error persists even at very high SNRs. © 2019 Association for Computing Machinery.

ACM International Conference Proceeding Series

Effect of sign-bit-flipping trojan on turbo coded communication systems

Due to high variability in shape, structure and occurrence; the non-palpable breast masses are often missed by the experienced radiologists. To aid them with more accurate identification, computer-aided detection (CAD) systems are widely used. Most of the developed CAD systems use complex handcrafted features which introduce difficulties for further improvement in performance. Deep or high-level features extracted using deep learning models already have proven its superiority over the low or middle-level handcrafted features. In this paper, we propose an automated deep CAD system performing both the functions: mass detection and classification. Our proposed framework is composed of three cascaded structures: suspicious region identification, mass/no-mass detection and mass classification. To detect the suspicious regions in a breast mammogram, we have used a deep hierarchical mass prediction network. Then we take a decision on whether the predicted lesions contain any abnormal masses using CNN high-level features from the augmented intensity and wavelet features. Afterwards, the mass classification is carried out only for abnormal cases with the same CNN structure. The whole process of breast mass classification including the extraction of wavelet features is automated in this work. We have tested our proposed model on widely used DDSM and INbreast databases in which mass prediction network has achieved the sensitivity of 0.94 and 0.96 followed by a mass/no-mass detection with the area under the curve (AUC) of 0.9976 and 0.9922 respectively on receiver operating characteristic (ROC) curve. Finally, the classification network has obtained an accuracy of 98.05% in DDSM and 98.14% in INbreast database which we believe is the best reported so far. © 2019 IEEE.

Proceedings - 2019 IEEE International Conference on Data Science and Advanced Analytics, DSAA 2019

Towards automated breast mass classification using deep learning framework

In multicarrier scenarios it is important to maintain the signals power in the specified dynamic range for the QOS. Further in Digital Video Broadcasting-S2 (DVBS2) the variable coding and modulations are selected based on the individual channel power. The paper describes the power estimation algorithm used for monitoring of up linking signal power for onboard satellite payload under such scenarios. In this paper a comparison of the parametric and nonparametric approaches is done to find the best suitable algorithm for onboard application. The performance of each algorithm is verified and compared. The periodogram method is finalize due to low complexity. Finally, result of the performance of the selected periodogram method is reported under noisy and no noise condition. © 2019, Springer Nature Singapore Pte Ltd.

Advances in Intelligent Systems and Computing

Channel Power Estimation for DVBRCS to DVBS2 Onboard DSP Payload

Chemiluminescence is one of the most commonly used optical diagnostic techniques in combustion research where a line-of-sight projected information is generated from spatial fields. The exactness and uniqueness of reconstruction along with ease of implementation gives Abel inversion an edge over the other existing single-view reconstruction techniques for efficient estimation of spatial field from line-of-sight projections. Though there exist many such algorithms, the primary focus of these has been to ensure tractable inversion through a systematic regularization by imposing a smoothness constraint on discrete data points. But these techniques do not have the provision to process the input image prior to deconvolution in order to prevent accumulation of noise infiltrated during data acquisition. Another major limitation of these algorithms is to adopt the changes in characteristics of the input data points while maintaining optimal storage and time complexity. To address these issues, we have proposed a new image processing technique using standard Abel inversion for the application in combustion research. It provides a suitable model to ensure regularized inversion by imposing a smoothness constraint on acquired raw data. The new algorithm has been implemented to yield the physically significant chemiluminescence emission from hydroxyl radicals in flames from line-of-sight integrated images. The effectiveness of this algorithm is highlighted using exemplary OH chemiluminescence images captured from a standard swirl stabilized research burner. © 2019, Indian Academy of Sciences.

Fulltext

Sadhana - Academy Proceedings in Engineering Sciences

Application of image enhancement and mixture of Gaussian approach in combustion research

Unmanned Aerial Vehicles (UAVs) or drones are often used to reach remote areas or regions which are inaccessible to humans. Equipped with a large field of view, compact size, and remote control abilities, drones are deemed suitable for monitoring crowded or disaster-hit areas, and performing aerial surveillance. While research has focused on area monitoring, object detection and tracking, limited attention has been given to person identification, especially face recognition, using drones. This research presents a novel large-scale drone dataset, DroneSURF: Drone Surveillance of Faces, in order to facilitate research for face recognition. The dataset contains 200 videos of 58 subjects, captured across 411K frames, having over 786K face annotations. The proposed dataset demonstrates variations across two surveillance use cases: (i) active and (ii) passive, two locations, and two acquisition times. DroneSURF encapsulates challenges due to the effect of motion, variations in pose, illumination, background, altitude, and resolution, especially due to the large and varying distance between the drone and the subjects. This research presents a detailed description of the proposed DroneSURF dataset, along with information regarding the data distribution, protocols for evaluation, and baseline results. © 2019 IEEE.