Ananya Sen Gupta

Prof. Ananya Sen Gupta

orcid

Research Focus & Metrics

Academic Career Length

23Years (2004-Present)

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

Representative Publications

View All Publications
1 October 2020 in The Journal of the Acoustical Society of America
Verified
Journal Article
Closed access
PDF file
BKBernice Kubicek
Bernice Kubicek
AGAnanya Sen Gupta
Ananya Sen Gupta
IKIvars Kirsteins
Ivars Kirsteins
This paper introduces a feature extraction technique that identifies highly informative features from sonar magnitude spectra for automated target classification. The approach involves creating feature representations through convolution of a two-dimensional Gabor wavelet and acoustic color magnitudes to capture elastic waves. This feature representation contains extracted localized features in the form of Gabor stripes, which are representative of unique targets and are invariant of target aspect angle. Further processing removes non-informative features through a threshold-based culling. This paper presents an approach that begins connecting model-based domain knowledge with machine learning techniques to allow interpretation of the extracted features while simultaneously enabling robust target classification. The relative performance of three supervised machine learning classifiers, specifically a support vector machine, random forest, and feed-forward neural network are used to quantitatively demonstrate the representations' informationally rich extracted features. Classifiers are trained and tested with acoustic color spectrograms and features extracted using the algorithm, interpreted as stripes, from two public domain field datasets. An increase in classification performance is generally seen, with the largest being a 47% increase from the random forest tree trained on the 1–31 kHz PondEx10 data, suggesting relatively small datasets can achieve high classification accuracy if model-cognizant feature extraction is utilized.

1 November 2022 in The Journal of the Acoustical Society of America
Verified
Journal Article
Open access
PDF file
BKBernice Kubicek
Bernice Kubicek
AGAnanya Sen Gupta
Ananya Sen Gupta
IKIvars Kirsteins
Ivars Kirsteins
Sonar target recognition remains an active area of research due to the complex entanglement of features from various acoustic scatterers, background clutter, and distortion by waveguide propagation effects. An equally challenging issue is due to different acoustic echoes returned from the target (including different target elements) itself. This work investigates the sonar target classification problem from a statistical perspective and aims to extract salient target feature vectors. Specifically, a multivariate statistical method is employed, canonical correlation analysis (CCA), as a feature extraction technique prior to multi-class classification of active sonar field data. The intuition behind using CCA is that persistent features slowly morph over time due to the changing aspect angles and platform positions and can be represented by maximally correlated projections of consecutive pings. CCA is applied using a sliding window, and the projections are used as feature vectors to train a neural network classifier. The smallest increase in classification accuracy when comparing the projection feature vectors to unprocessed feature vectors was 10%. The largest increase was 34%. The results are further examined through the use of confusion matrices and layer-wise relevance propagation, which distributes the trained networks output score to the input layer.

28 November 2016 in BMC Chemistry
Verified
Journal Article
Open access
PDF file
HDHamidreza Ghasemi Damavandi
Hamidreza Ghasemi Damavandi
AGAnanya Sen Gupta
Ananya Sen Gupta
RNRobert K. Nelson
Robert K. Nelson
CRChristopher Reddy
Christopher Reddy
Comprehensive two-dimensional gas chromatography $$(GC \times GC)$$ provides high-resolution separations across hundreds of compounds in a complex mixture, thus unlocking unprecedented information for intricate quantitative interpretation. We exploit this compound diversity across the $$(GC \times GC)$$ topography to provide quantitative compound-cognizant interpretation beyond target compound analysis with petroleum forensics as a practical application. We focus on the $$(GC \times GC)$$ topography of biomarker hydrocarbons, hopanes and steranes, as they are generally recalcitrant to weathering. We introduce peak topography maps (PTM) and topography partitioning techniques that consider a notably broader and more diverse range of target and non-target biomarker compounds compared to traditional approaches that consider approximately 20 biomarker ratios. Specifically, we consider a range of 33–154 target and non-target biomarkers with highest-to-lowest peak ratio within an injection ranging from 4.86 to 19.6 (precise numbers depend on biomarker diversity of individual injections). We also provide a robust quantitative measure for directly determining “match” between samples, without necessitating training data sets. We validate our methods across 34 $$(GC \times GC)$$ injections from a diverse portfolio of petroleum sources, and provide quantitative comparison of performance against established statistical methods such as principal components analysis (PCA). Our data set includes a wide range of samples collected following the 2010 Deepwater Horizon disaster that released approximately 160 million gallons of crude oil from the Macondo well (MW). Samples that were clearly collected following this disaster exhibit statistically significant match $$(99.23 \pm 1.66 )\,\%$$ using PTM-based interpretation against other closely related sources. PTM-based interpretation also provides higher differentiation between closely correlated but distinct sources than obtained using PCA-based statistical comparisons. In addition to results based on this experimental field data, we also provide extentive perturbation analysis of the PTM method over numerical simulations that introduce random variability of peak locations over the $$(GC \times GC)$$ biomarker ROI image of the MW pre-spill sample (sample $$\#1$$ in Additional file 4: Table S1). We compare the robustness of the cross-PTM score against peak location variability in both dimensions and compare the results against PCA analysis over the same set of simulated images. Detailed description of the simulation experiment and discussion of results are provided in Additional file 1: Section S8. We provide a peak-cognizant informational framework for quantitative interpretation of $$(GC \times GC)$$ topography. Proposed topographic analysis enables $$(GC \times GC)$$ forensic interpretation across target petroleum biomarkers, while including the nuances of lesser-known non-target biomarkers clustered around the target peaks. This allows potential discovery of hitherto unknown connections between target and non-target biomarkers.

24 August 2023 in IEEE Sensors Letters
Verified
Journal Article
Closed access
AGAnanya Sen Gupta
Ananya Sen Gupta
DMDavid Miles
David Miles
Magnetic interference from the spinning reaction wheels often used to control spacecraft orientation is a well-known issue for accurate sensing of geophysical measurements using sensitive magnetometers that capture the geophysical magnetic fields in the macroscale (order of 1 kilometers or greater). Disentangling the reaction wheel interference from the background measurements has been an open challenge to geomagnetic signal analysis for several decades. In this work, we address this challenge as a feature separation problem, and leverage the informational redundancy and feature continuity across multiple spatial dimensions and dual sensors to calculate accurate spectral support estimates of the reaction wheel signature. Specifically, we exploit the feature continuity across multiple spectrograms to detect the multi-harmonic reaction wheel signature across multiple spatial dimensions and inboard and outboard sensor measurements against the background geophysical magnetic field. A representative case study for the e-POP/Swarm- ECHO mission are presented.

28 December 2017 in Journal of Geophysical Research: Space Physics
Verified
Journal Article
Closed access
PDF file
AGAnanya Sen Gupta
Ananya Sen Gupta
CKCraig Kletzing
Craig Kletzing
RHRobin Howk
Robin Howk
WKWilliam Kurth
William Kurth
MMMorgan Matheny
Morgan Matheny

Verified
Journal Article
Open access
PDF file
BKBernice Kubicek
Bernice Kubicek
AGAnanya Sen Gupta
Ananya Sen Gupta
IKIvars Kirsteins
Ivars Kirsteins
Active sonar target classification remains an ongoing area of research due to the unique challenges associated with the problem (unknown target parameters, dynamic oceanic environment, different scattering mechanisms, etc.). Many feature extraction and classification techniques have been proposed, but there remains a need to relate and explain the classifier results in the physical domain. This work examines convolutional neural networks trained on simulated data with a known ground truth projected onto two time-frequency representations (spectrograms and scalograms). The classifiers were trained to discriminate the target material type, geometry, and internal fluid filling, while the hyperparameters were tuned to the classification task using Bayesian optimization. The trained networks were examined using an explainable artificial intelligence technique, gradient-weighted class activation mapping, to uncover the informative features used in discrimination. This analysis resulted in visual representations that allowed the CNN choices to be related to the physical domain. It was found that the scalogram representation provided a negligible classification accuracy increase compared with the spectrograms. Networks trained to discriminate between target geometries resulted in the highest accuracy, and the networks trained to discriminate the internal fluid of the target resulted in the lowest accuracy.