Sowmitra Das

Sowmitra Das

Auto-generated profile, based on publicly available information

Sowmitra Das

Department of Theoretical Physics, University of Dhaka, Dhaka, Bangladesh

Research Focus & Metrics

Academic Career Length

16Years (2011-Present)

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

Representative Publications

View All Publications
26 April 2019 in IEEE Transactions on Intelligent Transportation Systems
Verified
Journal Article
Closed access
UKUday Kamal
Uday Kamal
TTThamidul Islam Tonmoy
Thamidul Islam Tonmoy
SDSowmitra Das
Sowmitra Das
KHKamrul Hasan
Kamrul Hasan
Traffic sign detection is a central part of autonomous vehicle technology. Recent advances in deep learning algorithms have motivated researchers to use neural networks to perform this task. In this paper, we look at traffic sign detection as an image segmentation problem and propose a deep convolutional neural network-based approach to solve it. To this end, we propose a new network, the SegU-Net, which we form by merging the state-of-the-art segmentation architectures-SegNet and U-Net to detect traffic signs from video sequences. For training the network, we use the Tversky loss function constrained by an L1 term instead of the intersection over union loss traditionally used to train segmentation networks. We use a separate network, inspired by the VGG-16 architecture, to classify the detected signs. The networks are trained on the challenge free sequences of the CURE-TSD dataset. Our proposed network outperforms the state-of-the-art object detection networks, such as the Faster R-CNN inception Resnet V2 and R-FCN Resnet 101, by a large margin and obtains a precision and recall of 94.60% and 80.21%, respectively, which is the current state of the art on this part of the dataset. In addition, the network is tested on the German Traffic Sign Detection Benchmark (GTSDB) dataset, where it achieves a precision and recall of 95.29% and 89.01%, respectively. This is on a par with the performance of the aforementioned object detection networks. These results prove the generalizability of the proposed architecture and its suitability for robust traffic sign detection in autonomous vehicles.

22 July 2021 in Quantum Information Processing
Verified
Journal Article
Closed access
PDF file
SDSowmitra Das
Sowmitra Das
SRSaifur Rahman
Saifur Rahman
MMMahbub Majumdar
Mahbub Majumdar

1 December 2011
Verified
Conference Paper
Closed access
HSHanif Seddiqui
Hanif Seddiqui
SDSowmitra Das
Sowmitra Das
IAIqbal Ahmed
Iqbal Ahmed
RNRudra Pratap Deb Nath
Rudra Pratap Deb Nath
MAMasaki Aono
Masaki Aono
The proliferation of heterogeneous sources of ontology instances of semantic knowledge base raises a research issue of automatic matching of instances. However, automatic instance matching is heavily affected by the weight of property associated to instances. Measuring the property weight automatically is a formidable task. In this paper, we propose an efficient method of measuring weight automatically and apply the method for augmentation of our state-of-the-art instance matcher, which consider the semantic specification of properties associated to instances, for matching with heterogeneous instances of semantic knowledge base. Our experiments and evaluations shows the effectiveness of automatic weight generation in ontology instance matching over various transformations of a dataset: value transformation, logical and structural transformation.

1 December 2018
Verified
Conference Paper
Closed access
SDSowmitra Das
Sowmitra Das
BSBejoy Sikder
Bejoy Sikder
HKHadiur Rahman Khan
Hadiur Rahman Khan
TITariqul Islam
Tariqul Islam
In this paper, we present various microstrip-line fed gap-coupled antennas for different patch-geometries. A single shorting post with a pair of λ / 4 resonator and a radiating patch have been used in the antenna to obtain improved Voltage Standing Wave Ratio (VSWR). We have proposed 5 different patch-geometries - square, rectangle, pentagon, trigonal pentagon, five-spoked star - which exhibit superior bandwidth and VSWR properties in the 4-7 GHz frequency range. The proposed antennas occupy a small space and have been simulated on a 1.6 mm thick ArIon Substrate. The (VSWR) for the five proposed patch-geometries have been simulated, and, their radiation patterns have been compared across the VSWR ≤ 2 bandwidth.