Publication

May 10, 2020

Touchless Diagnostics using Artificial Intelligence Algorithms and Deep Learning

The philosophical foundation of the paper is on algorithmic applications on human data collected via non-evasive touchless fashion to generate medical insights which can be used as a template or recommendation or triage factor for medical professionals by acting as Diagnostic Decision Support system for doctors. All these algorithms are all Machine Learning Algorithms using Artificial Neural Networks.
May 10, 2020

Machine Learning And hyperspectral Imaging on GIS data for Computer Vision Based Analysis of Agriculture Loans

Using Artificial Intelligence Algorithms on Satellite Data feeds to do real-time analysis of Quality and Quantity of Agricultural yields of a certain location. We use data from 3 separate satellites including HyperSpectral Imaging data analysis for this process. It is used to Automate agricultural loans.
May 10, 2020

Analysis of expression of luteal genes during induced luteolysis and rescue of luteal function in bonnet macaques and Pregnant Rats

Studies have been carried out to standardize induced luteolysis model systems ƵƟůŝnjŝnŐ female monkeys and pregnant rats. In monkeys, ĂĚmŝnŝƐƚrĂƟŽn of a single ŝnũĞcƟŽn of GnRH receptor antagonist, Cetrorelix (CET; 150 µg/kg BW s.c.,), on day 7 of the luteal phase led to profound decrease in serum progesterone (P4) cŽncĞnƚrĂƟŽn within 24 h (3.6 ± 1.1 vs 0.8 ± 0.2 ng/ml before and 24 h post CET, rĞƐƉĞcƟvĞůy p<0.05), and followed by the premature onset of menses 96 h later.
May 10, 2020

QUICKSAL A small and sparse visual saliency model for efficient inference in resource constrained hardware

Visual saliency is an important problem in the field of cognitive science and computer vision with applications such as surveillance, adaptive compressing, detecting unknown objects and scene understanding. In this paper, we propose a small and sparse neural network model for performing salient object segmentation that is suitable for use in mobile and embedded applications. Our model is built using depthwise separable convolutions and bottleneck inverted residuals which have been proven to perform very memory efficient inference and can be easily implemented using standard functions available in all deep learning frameworks.