Projects
Emotion Recognition Using Body Gesture
(January 2020 - December 2021)
Body Gesture is an important way of nonverbal communication, specially in the era of covid-19 where most of the people cover up their faces wearing masks. My research focuses on recognizing six emotions - Anger, Fear, Happiness, Sadness, Disgust, Surprise and Neutral, using deep learning algorithms, from body gesture in real time.
To learn more about the project visit: https://github.com/Tamima0049/Emotion-from-Body-Language
Applications:
The goal of this research is to help the children with Autism Spectrum Disorder to recognize emotion in people using this app in real time.
​
Image processing GUI with spatial and frequency domain filters using MATLAB:
(Mar 2020 – May 2020)
In this project I implemented different filtering techniques on images. I combined all these filters in a GUI app. In this project I learnt how to build a graphical user interface app using Matlab. I learnt the significance of different types of image processing filters and worked with their applications. I came to see the difference between the spatial domain filters and frequency domain filters. I learnt when I need which filter depending on the image. From this project the users can get a ready app where they can upload their desired images, apply different filters on the images and save it back to the desired computer location. To use this app, the user should have minimal image processing knowledge such as the use of appropriate cutoff frequency, filter radius, order of the filter etc.
Basic statistical distribution using python
(Nov 2019 – Dec 2019)
In this particular study different types of statistical distributions (normal, uniform, poisson, binomial, exponential and t distribution) were solved with Python programming language. Python library has all the necessary tools to calculate the probabilities and visualize the distributions. Different functions - such as pmf, cdf, etc, associated with these distributions can also be calculated easily with python. In this study the use of these python statistical problem-solving tools were shown. These python tools provide exact solutions of the distribution problems which is much time consuming if calculated manually. The jupyter computing platform was used for this project.
(June 2021 - December 2021)
The HiPE group has been sponsored and partnered with Dell Technologies to perform a Hardware Benchmarking Performance Analysis Using Machine Learning Models to their AI Data Science Workstations (DSW) and mobile solutions . I worked as a research assistant for this benchmarking project. The research includes benchmarking of three of their work stations- Precision 7750 (NVIDIA® Quadro RTX™ 5000 graphic card), Tower 5820 (NVIDIA® Quadro RTX™ 6000 graphic card) and Tower 7920 (NVIDIA® Quadro RTX™ 6000 graphic card). The goal was to compare the performance of these three work stations and their existing data-science-stack environments to other existing resources of HiPE and Texas State University. Also, Multithreading and multiprocessing technique were used to leverage the multicore workstations and improve the models' efficiency.
Optimizing A Bank Model Capacity using Forecasting and Simulation:
(April 2021- May 2021)
In this project, we modeled a bank system consisting of a single Sign-In employee, three ATM machines, four Tellers, and two Managers. Customers who visited the bank were categorized into three groups based on the services they were directed to by the Sign-In employee – Basic, Moderate and Advanced. A spreadsheet containing data of 500 customers was provided which contains the type of customers entering the bank, the Interarrival times for all customer types, the Sign-In times for all customer types, the service each customer was directed to, and the Service times of the ATMs, Tellers, and Managers. The bank was planning on hiring two additional employees to improve the system, and it was our goal to use simulation and find out which service(s) would yield the best system performance when provided these additional employees.
Magnetic loop antenna design using Hfss:
(Oct 2019 – Dec 2019)
An efficient magnetic loop antenna was designed using HFSS (High Frequency Structure Simulator). Different parameters of the antenna such directivity and S11 parameters were analyzed. The S11 parameter was around 5.9 GHZ and the antenna gain was in negative dB values. So, further investigation was needed on its surrounding features for the directivity to be increased. Also, different parameters of the antenna were changed and then antenna performance were analyzed based on the changes of their parameters.
