Ankur Shrivastava

Master Thesis : Exploring Deep Clustering features for automatical optical detection. *Adaptation of unsupervised method of Deep clustering (neural network with K-mean clustering) *Comparing Deep clustering features with supervised deep learning features for image retrieval and anoamly detection *Used Deep learning framework Caffe on MATLAB and SciKit learning for clustering in python.

Intern : Convolution Neural Network for defect detection in automatical optical detection system. *Explored various preprocessing methods and CNN-architecture (ResNet-100, ResNet-50) for accuracy improvement. *Tried various methods for finding similar image from a database using CNN features example Euclidian distance, extreme value signature, cosine distance. *Used Deep learning framework Caffe on MATLAB and python.

*Managing and developing C++ codes for database management using integrated SQL scripts. *Developed a payment gateway module for an American client in C++. *Daily production and support handling using ‘Cribmaster’ tool. *Unix shell scripting for managing large database

* CGPA - 1.5 * Image Processing * Deep Learning * Machine Learning * Signal Processing * Speech Recognition

* Electrical Engineering * Image Processing * C/C**