Rajiv Bagai

Dr. Rajiv Bagai

School of Computing, Wichita State University, Wichita, Kansas, USA
orcid

Research Focus & Metrics

Academic Career Length

15Years (2012-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
16 November 2012 in IEEE Systems Journal
Verified
Journal Article
Closed access
Rajiv BagaiRajiv Bagai
Rajiv Bagai
BTBin Tang
Bin Tang
EKEuna Kim
Euna Kim
A major objective of any system-wide attack on an anonymity system is to uncover the extent to which each user of the system communicated with each other user. A probabilistic attack attempts to achieve this objective by arriving at some probability values for each of the system's possible input–output message pairs of reflecting actual communication. We show that these values lead to a probability distribution on the set of all possible system-wide communication patterns between users, and develop a combinatorial technique to determine this distribution. We give a method to measure from this distribution the effectiveness of any such attack or, alternatively, the level of anonymity remaining in the system in the aftermath of the attack. We also compare our metric with three earlier attempts in the literature to solve a similar problem, and demonstrate that the scope of our metric is far wider than those of all earlier ones.

20 January 2017 in IEEE Transactions on Information Forensics and Security
Verified
Journal Article
Closed access
Rajiv BagaiRajiv Bagai
Rajiv Bagai
NMNafia Malik
Nafia Malik
MJMurtuza Jadliwala
Murtuza Jadliwala
There is clear demand among organizations for sharing their data for mining and other purposes without compromising the privacy of individual objects contained in the data. Pseudonymization is a simple, yet widely employed technique for sanitizing such data prior to its release; it replaces identifying names in the data by pseudonyms. Well-known metrics already exist in the literature for measuring the amount of anonymity still contained in some pseudonymized data in the aftermath of an infeasibility background attack. While the need for a metric for the much wider and more realistic class of probabilistic background attacks has also been well identified, currently no such metric exists. We fulfill that long identified need by presenting two metrics, an approximate and a more exact one, for measuring anonymity in pseudonymized data in the wake of a probabilistic attack. These metrics are rather intractable, thus impractical to employ in real-life situations. Therefore, we also develop an efficient heuristic for our superior metric, and show the remarkable accuracy of our heuristic. Our metrics and heuristic assist a data owner in evaluating the safety level of pseudonymized data against probabilistic attacks before making a decision on its release.

12 May 2026 in Remote Sensing
Verified
Journal Article
Open access
PDF file
SKShruti Rajendra Kshirsagar
Shruti Rajendra Kshirsagar
BCBharath Chandra
Bharath Chandra
UTUnaza Tallal
Unaza Tallal
Rajiv BagaiRajiv Bagai
Rajiv Bagai
ADAtri Dutta
Atri Dutta
Automated building damage assessment from satellite imagery has become increasingly critical for rapid disaster response and humanitarian relief operations. However, current state-of-the-art deep learning models exhibit significant generalization challenges when deployed to geographically and environmentally diverse regions. This study investigates the nature and extent of geographic bias in building damage detection systems, revealing that model performance degradation occurs primarily from geographic and structural characteristics rather than insufficient training data representation. Through a systematic evaluation of top-performing xView2 competition solutions across 17 disaster locations across multiple climate zones, we found that even state-of-the-art models struggle with generalization under geographic shift, particularly for the minor and major damage classes, and exhibit strong geographic biases toward certain regions. In this work, we explore a six-channel Fusion Augmentation strategy that enriches RGB imagery with auxiliary structural enhancement channels, together with supervised fine-tuning and unsupervised CORAL-based domain adaptation for three unseen regions. The experimental results demonstrate a substantial improvement of 7.1% overall F1 score, with notable gains for intermediate damage categories such as minor and major damage. Domain adaptation experiments on three unseen locations show that combining Fusion Augmentation with supervised fine-tuning yields 40.8% and 60.0% improvements over the minor and major classes, while unsupervised CORAL achieves 24.2% and 39.5% improvements over the minor and major damage classes compared to benchmarks. These findings highlight persistent geographic bias and demonstrate that structural feature enhancement combined with domain adaptation is essential for globally deployable damage assessment systems.

1 June 2016
Verified
Book Chapter
Closed access
PDF file
Rajiv BagaiRajiv Bagai
Rajiv Bagai
HLHuabo Lu
Huabo Lu


26 March 2026 in Preprints.org
Verified
Preprint
Open access
PDF file
SKShruti Rajendra Kshirsagar
Shruti Rajendra Kshirsagar
BCBharath Chandra
Bharath Chandra
UTUnaza Tallal
Unaza Tallal
Rajiv BagaiRajiv Bagai
Rajiv Bagai
ADAtri Dutta
Atri Dutta
Automated building damage assessment from satellite imagery has become increasingly critical for rapid disaster response and humanitarian relief operations. However, current state-of-the-art deep learning models exhibit significant generalization challenges when deployed to geographically and environmentally diverse regions. This study investigates the nature and extent of geographic bias in building damage detection systems, revealing that model performance degradation stems primarily from geographic and structural characteristics rather than insufficient training data representation. Through systematic evaluation of top-performing xView2 competition solutions across 17 disaster locations spanning multiple climate zones, we found that even state-of-the-art models struggle with generalization, particularly for Minor and Major damage classes, and exhibit strong geographic biases toward certain regions. Strikingly, Nepal despite having the largest training dataset (15,234 images) achieves the worst performance, demonstrating that geographic and structural characteristics dominate generalization behavior more than training data quantity. To address these fundamental limitations, we explore Fusion Augmentation, a novel methodology that enhances edge detection and structural feature representation by integrating auxiliary information channels with standard RGB imagery. Experimental results demonstrate substantial improvements of 7.1% overall F1 score, with dramatic gains for intermediate damage categories such as Minor and Major damage. Domain adaptation experiments on three unseen locations show that combining Fusion Augmentation with supervised fine-tuning yields 40.8% and 60.0% improvements over Minor and major classes, while unsupervised CORAL achieves 24.2% and 39.5% improvements over Minor and major damage classes compared to benchmarks. These findings challenge prevailing assumptions about data-driven generalization in remote sensing AI systems and demonstrate that structural feature enhancement combined with domain adaptation is essential for robust detection across geographically diverse deployment scenarios, providing practical strategies for globally deployable damage assessment systems.