PRIYASH BHUGRA

- Multi-temporal and mono-temporal analysis of geospatial images for multi-label classification. - LULC classification and change detection of geospatial images. - Preparation of artificial satellite images for change detection using Adversarial neural network - LULC maps visualization using MAPBOX - Monitoring machine learning models for handling data and concept drift. - Pre-processing of geospatial images using gdal, rasterio

German Traffic sign detection using state of the art deep learning approaches. Point could reconstruction using Blender for Data Preparation. Role: - Developed deep learning models for small object detection - Annotated traffic sign for better prediction - Implemented different kinds of data augmentation techniques on Images - Achieved state of the art on German Traffic Sign Detection Benchmark (GTSDB) - Prepared dataset for point cloud segmentation using Blender

Multi-Pedestrian Tracking by Integration of Particle Filter with Faster-RCNN for Autonomous Vehicle Role: - Developed a Deep Learning model for pedestrian tracking - Implemented Faster-RCNN (object detection model from scratch) - Implemented a tracker based on a Particle filter - Integrated Particle filter in FasterRCNN for better tracking metrics

Identifying if a person is Normal/Abnormal by analysing electroencephalogram (EEG) signals from Temple university dataset using Recurrent Neural Network (LSTM, GRU).

Projects: 1) Online Vehicle Booking System Technology: Java, MVC Architecture, J2EE technologies (JSP, Servlets), SQL. Role: Web Developer 2) Identity and Access Management. Technology: Java, SQL, OIM, OAM, OID. Client: SBI General Insurance, Nomura Bank. Role: IAM Analyst

Specialized in Artificial intelligence and software engineering