Computer Vision/Machine Learning Engineer
Resume:
Abinaya Manimaran
̋ www.linkedin.com/in/B abinaya-manimara@manimaran/ usc.edu
https://github.com/mcabinaya/
EDUCATION
Aug 2017 -
May 2019
Master of Science, Electrical Engineering (Computer Vision and Machine Learning) GPA: 3.94/4.0, University of Southern California, Los Angeles, California, USA, Relevant Coursework: EE 660 - Machine Learning From Signals: Foundations and Methods, EE 500 - Neural Learning and Computational Intelligence, EE 569 - Digital Image Processing, EE 503 - Probability for Electrical and Computer Engineers, EE 559 - Mathematical Pattern Recognition, EE 510 - Linear Algebra for Engineering, EE 511 - Simulation Methods for Stochastic Systems. Jun 2010 -
May 2014
Bachelor Of Engineering, Electronics And Communication Engineering GPA: 9.76/10.0, Thiagarajar College Of Engineering, Anna University, Madurai, India Rank: 1/150. EXPERIENCE
Aug 2018 -
Present
Journeyman Fellow, U.S. Army Research Laboratory, Los Angeles (ARL West). Research Areas: Adversarial Attacks, SAAK Transform, Deep Neural Networks, Computer Vision [TensorFlow, Keras]
{ A novel lossy SAAK Transform is proposed as a pre-processing tool to defend against adversarial attacks in DNNs
{ Several experiments being conducted to extend this work on many benchmark image datasets
{ Proposed research work will be submitted to Computer Vision and Pattern Recognition (CVPR), 2019 Conference May 2018 -
Aug 2018
Computer Vision Research Intern, HERE Technologies, Berkeley. Research Areas: Super Resolution, Computer Vision, Deep Learning, Highly Automated Driving [PyTorch, CUDA]
{ Developed expertise in training Convolutional Neural Networks, Residual Networks and Generative Adversarial Networks
{ Proposed perceptual loss based super resolution model - using pre-trained network as loss function
{ Evaluated our task based loss function for SR-ResNet by human grading Oct 2017 -
Aug 2018
Graduate Research Assistant, Innovation in Integrated Informatics Lab (i-Lab), USC. Research Areas: Time Series Analysis, Pattern Recognition, Unsupervised Learning, Sensor Data [Python]
{ Project: Smarter Environments - "Life of a Desk", a collaboration between ARUP IoT and USC
- Occupancy prediction and profiling from motion sensor data
- Laptop Usage Prediction and profiling from power consumption data July 2014 -
July 2017
Researcher, TCS Innovation Labs (Tata Consultancy Services Ltd.), Chennai. Research Areas: Statistical Signal Processing, Machine Learning, Cyber-Physical Systems [Python, R]
{ Improved pipe burst prediction accuracy in water utility networks by combining data driven and domain based knowledge
{ Gained hands-on experience training machine learning models - Naive Bayes, Logistic Regression, Support vector Machines, Artificial Neural Networks, Rankboost etc.
- Paper published in IAAI-17 (https://aaai.org/ocs/index.php/IAAI/IAAI17/paper/view/14356/)
- Patent filed under Europe, India and USA regions Jan 2014-
May 2014
Research Intern, TCS Innovation Labs (Tata Consultancy Services Ltd.), Chennai. Research Areas: Cyber-Physical Systems, Time Series Analysis, Pattern Recognition, Anomaly detection [Python]
{ Analyzed energy consumption in retail stores to identify energy saving opportunities
{ Modeled energy data using ARIMA analysis and Principal Component Analysis SKILLS
Languages Python (Numpy, Pandas, Scikit-Learn, Matplotlib, OpenCV, TensorFlow, Keras, PyTorch), C++, R Others MATLAB, Weka, GIS, GeoServer, LATEX, EPANET, NI LabVIEW OS Windows, GNU Linux (Ubuntu), macOS
PUBLICATIONS
2017 Gollakota Kaushik*, Abinaya Manimaran*, Arunchandar Vasan, Venkatesh Sarangan and Anand Sivasubramaniam: Burst Prediction in Water Networks Using Dynamic Metrics. Proceedings of the 29th Annual Conference on Innovative Applications of Artificial Intelligence (IAAI-17)
PROJECTS
2018 Digital Image Processing [C++, OpenCV]
Image Denoising, Creating Special Effects, Geometric Image Modification, Homographic Transformation, Morphological Processing, Texture Classification and Segmentation, Edge Detection, Image Matching Using SIFT/SURF, MNIST Classification/Comparison Using LeNet-CNN and Subspace Approximation and Augmented Kernels (SAAK) Transform. 2018 Bank Marketing Data Analysis [Python, Scikit-Learn] Explored supervised models to analyze and predict if a client will subscribe for a term deposit given his/her marketing campaign related data. Pre-processed dataset with missing values, handled class imbalance, and best model was fine tuned for hyper parameters by k-fold Cross Validation.
2018 Simulation Methods For Stochastic Systems [Python] Double Rejection, K-Means Clustering, Gaussian Mixture Models, Bag of Words Clustering, Monte Carlo Approximation and Integration, Markov Chain Monte Carlo (MCMC) Based Sampling, Optimization and Traveling Salesman Problem 2013 Topological Texture Feature Based Age Classification [MATLAB] Exploited different feature extraction methods to train a K-Nearest Neighbor classifier.The proposed methodology, was able to classify facial images into 6 age groups with 73 percentage of accuracy. 2013 Detection and Removal of Poisson noise in Low Light Images [MATLAB] Noise distribution was analyzed to set a threshold and identify images as noisy or noiseless. Transformation techniques were used to effectively denoise images.
2012 Weapon Detection in X-ray Images using Bayes classifier [MATLAB] Textural features from x-ray weapon images were extracted. Bayes Classifier was used to give timely alert of a concealed weapon and identify its location on the image of baggage.
* indicates joint first authors
Contact this candidate