Manish Sharma

Dr. Manish Sharma


Research Focus & Metrics

Academic Career Length

7Years (2020-Present)

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

Representative Publications

View All Publications
30 November 2020 in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Verified
Journal Article
Open access
MSManish Sharma
Manish Sharma
MDMayur Dhanaraj
Mayur Dhanaraj
SKSrivallabha Karnam
Srivallabha Karnam
DCDimitris G. Chachlakis
Dimitris G. Chachlakis
RPRaymond Ptucha
Raymond Ptucha
ESEli Saber
Eli Saber
Deep-learning object detection methods that are designed for computer vision applications tend to under-perform when applied to remote sensing data. This is because, contrary to computer vision, in remote sensing training data are harder to collect and targets can be very small, occupying only a few pixels in the entire image, and exhibit arbitrary perspective transformations. Detection performance can improve by fusing data from multiple remote sensing modalities, including RGB, IR, hyper-spectral, multi-spectral, synthetic aperture radar, and LiDAR, to name a few. In this work, we propose YOLOrs: a new convolutional neural network, specifically designed for realtime object detection in multimodal remote sensing imagery. YOLOrs can detect objects at multiple scales, with smaller receptive fields to account for small targets, as well as predict target orientations. In addition, YOLOrs introduces a novel midlevel fusion architecture that renders it applicable to multimodal aerial imagery. Our experimental studies compare YOLOrs with contemporary alternatives and corroborate its merits.

28 September 2023 in arXiv
Verified
Preprint
Open access
PDF file
MSManish Sharma
Manish Sharma
MCMoitreya Chatterjee
Moitreya Chatterjee
KPKuan-Chuan Peng
Kuan-Chuan Peng
SLSuhas Lohit
Suhas Lohit
MJMichael Jones
Michael Jones
The primary bottleneck towards obtaining good recognition performance in IR images is the lack of sufficient labeled training data, owing to the cost of acquiring such data. Realizing that object detection methods for the RGB modality are quite robust (at least for some commonplace classes, like person, car, etc.), thanks to the giant training sets that exist, in this work we seek to leverage cues from the RGB modality to scale object detectors to the IR modality, while preserving model performance in the RGB modality. At the core of our method, is a novel tensor decomposition method called TensorFact which splits the convolution kernels of a layer of a Convolutional Neural Network (CNN) into low-rank factor matrices, with fewer parameters than the original CNN. We first pretrain these factor matrices on the RGB modality, for which plenty of training data are assumed to exist and then augment only a few trainable parameters for training on the IR modality to avoid over-fitting, while encouraging them to capture complementary cues from those trained only on the RGB modality. We validate our approach empirically by first assessing how well our TensorFact decomposed network performs at the task of detecting objects in RGB images vis-a-vis the original network and then look at how well it adapts to IR images of the FLIR ADAS v1 dataset. For the latter, we train models under scenarios that pose challenges stemming from data paucity. From the experiments, we observe that: (i) TensorFact shows performance gains on RGB images; (ii) further, this pre-trained model, when fine-tuned, outperforms a standard state-of-the-art object detector on the FLIR ADAS v1 dataset by about 4% in terms of mAP 50 score.

16 January 2024 in arXiv
Verified
Preprint
Open access
PDF file
MSManish Sharma
Manish Sharma
JHJamison R Heard
Jamison R Heard
ESEli Saber
Eli Saber
PMPanos P. Markopoulos
Panos P. Markopoulos
While Convolutional Neural Networks (CNNs) excel at learning complex latent-space representations, their over-parameterization can lead to overfitting and reduced performance, particularly with limited data. This, alongside their high computational and memory demands, limits the applicability of CNNs for edge deployment. Low-rank matrix approximation has emerged as a promising approach to reduce CNN parameters, but its application presents challenges including rank selection and performance loss. To address these issues, we propose an efficient training method for CNN compression via dynamic parameter rank pruning. Our approach integrates efficient matrix factorization and novel regularization techniques, forming a robust framework for dynamic rank reduction and model compression. We use Singular Value Decomposition (SVD) to model low-rank convolutional filters and dense weight matrices and we achieve model compression by training the SVD factors with back-propagation in an end-to-end way. We evaluate our method on an array of modern CNNs, including ResNet-18, ResNet-20, and ResNet-32, and datasets like CIFAR-10, CIFAR-100, and ImageNet (2012), showcasing its applicability in computer vision. Our experiments show that the proposed method can yield substantial storage savings while maintaining or even enhancing classification performance.