A Case for PARAM Shavak: Ready-to-Use and Affordable Supercomputing Solution

Agrawal, Sandeep; Das, Shirshendu; Valmiki, Manjunatha; Wandhekar, Sanjay; Moona, Rajat

doi:10.1109/hpcs.2017.66

Cited by 7 publications

(2 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Videos of 640 9 480 pixels, 30 frames per second AVI format without audio have been used for conducting the experiments. Param Shavak [21] supercomputer has been used for training the proposed system model and MobaXterm has been used for accessing the same.…”

Section: Methodsmentioning

confidence: 99%

Deep CNN models-based ensemble approach to driver drowsiness detection

Dua

Shakshi

Singla

et al. 2020

Neural Comput & Applic

112

View full text Add to dashboard Cite

Statistics have shown that many accidents occur due to drowsy condition of drivers. In a study conducted by National Sleep Foundation, it has been found that about 20% of drivers feel drowsy during driving. These statistics paint a very scary picture. This paper proposes a system for driver drowsiness detection, in which the architecture detects sleepiness of driver. The proposed architecture consists of four deep learning models: AlexNet, VGG-FaceNet, FlowImageNet and ResNet, which use RGB videos of drivers as input and help in detecting drowsiness. Also, these models consider four types of different features such as hand gestures, facial expressions, behavioral features and head movements for the implementation. The AlexNet model is used for various background and environmental changes like indoor, outdoor, day and night. VGG-FaceNet is used to extract facial characteristics like gender ethnicities. FlowImageNet is used for behavioral features and head gestures, and ResNet is used for hand gestures. Hand gestures detection provides a precise and accurate result. These models classify these features into four classes: non-drowsiness, drowsiness with eye blinking, yawning and nodding. The output of these models is provided to ensemble algorithm to obtain a final output by putting them through a SoftMax classifier that gives us a positive (drowsy) or negative answer. The accuracy obtained from this system came out to be 85%.

show abstract

Section: Methodsmentioning

confidence: 99%

Deep CNN models-based ensemble approach to driver drowsiness detection

Dua

Shakshi

Singla

et al. 2020

Neural Comput & Applic

112

View full text Add to dashboard Cite

show abstract

“…To train our model we used C-DAC PARAM Shavak DL GPU System [23,26], which is equipped with the latest × 86 based Intel Xeon server series processor along with one Nvidia Quadro P5000 accelerator card. It has 64 GB of RAM and 8 TB of secondary storage.…”

Section: Training Detailsmentioning

confidence: 99%

Efficient and Robust Indian Number Plate Recognition Through Modified and Tuned LPRNet

Dangi

Kuvelkar

Maity

et al. 2021

Lecture Notes in Electrical Engineering

Self Cite

View full text Add to dashboard Cite

This paper presents a novel approach for Indian Number Plate Recognition using the Convolution Neural Network (CNN). We generated and augmented synthetic data in Indian standards with variant background and font. We also collected and labelled real data of Indian License plates to create a genuine Indian oriented dataset. We compared existing models like SVM, KNN and CRNN in terms of accuracy and inference time, and selected LPRNet. We first trained the existing LPRNet model, which is designed for Chinese License plates on synthetic data and secured 85% accuracy. We then modified the LPRNet model and secured 93% accuracy on the same synthetic data. We applied the post-processing technique of pattern recognition rules for Indian standards of License plates to further improve accuracy. The Proposed system achieved 95% accuracy after fine-tuning with real data. Our model is very lightweight and can easily be deployed on Nvidia-Jetson nano, Intel Neural Compute Stick.

show abstract