Summary
Overview
Work History
Education
Skills
Final Year Project : Application of YOLO for detection and text extraction of HSRP/ NON HSRP number plates in video using Google Colab (Final year Project 2021)
Certification
Languages
Personal Information
Websites
Timeline
Generic
NRIP GUPTA

NRIP GUPTA

Delhi

Summary

I am a computer science engineer currently working as a Graduate Engineer- Digital (AI/ML) with Mott MacDonald India. I am open to new opportunities in the ML/DL domain. I have been passionately pursuing this domain since my B.Tech final year project and have been working on such projects as a professional for the last two years. I have completed a ML/DL certification programme from IIT Delhi and have a B.Tech in Computer Science and Engineering form Manipal Institute of Technology, Manipal (July 2021).

Overview

3
3
years of professional experience
1
1
Certification

Work History

Graduate Engineer- Digital (AI/ML)

Mott MacDonald India
Noida
10.2021 - Current

Health & Safety Detection Model (May 2023)

Customization and modification of an AI detection model for detection of safety helmets and vest on workers on site. An existing YOLO based model was customized to suit the required detection parameters. These changes were implemented using python. Modifications were made to increase the detection precision to a desired minimum of 95%.

Ash Dieback Detection Project (September 2022)

The goal of this project was to identify Ash trees infected by Dieback disease (for Prediction Model). A Pascal VOC based graded detection model was being considered to scale the degree of damage caused by the disease to identify the ones which could fall on highways leading to accidents/damage to infra.

Bird Call Detection (June 2022)

An attempt to develop a model to identify birds from their calls. It involved data scrapping to procure data from an online open-source database to be stored in CSV format.

Pavement Health Check Model (April 2022)

To develop a Visual Assessment model for detection of pavement defects by 3D Annotation of planned videos to distinguish different pavement defects and segmenting them using CVAT

Bridge Defects Detection Project (December 2021)

Detection of various types of defects in existing bridges by annotation of defects using Django Labeller in Anaconda Evnvironment

Presto Project (December 2021)

Creation of API to extract and process live streaming data from the National UK Rail server. Involved extraction and interpretation of live streaming XML data through Python requests and processing it into a user-friendly tabular format.

Sierra Leone Secondary Education Improvement Project (October 2021)

This was a major project to digitise the Sierra Leone education system through ML/AI. It involved handling large data sets, data extraction, collection, preparation, cleaning/ processing, simulation of model, transforming raw handwritten data into uniform type written format.

Education

Certification - Machine Learning And Deep Learning

IIT Delhi
Delhi
04-2024

B.Tech - Computer Science And Engineering

Manipal Institute of Technology, MAHE
Manipal
07-2021

Higher Secondary -

Delhi Public School Dwarka
07-2017

Skills

Tools & Programming skills

  • Programming Languages: Python, C, C, Java
  • ML Frameworks: TensorFlow, Scikit-Learn
  • Data Manipulation and Analysis: Pandas, NumPy
  • Computer Vision and Image Processing: OpenCV, NumPy
  • Data Visualization: Matplotlib
  • IDEs: Jupyter Notebook, Visual Studio Code

Domain specific knowledge

AI, ML & Deep learning

  • Computer Vision, Object Detection
  • Data Preprocessing and Cleaning
  • Supervised and Unsupervised Learning
  • Neural Networks: CNNs, RNNs, LSTM, Autoencoder
  • YOLO algorithm (deep learning)
  • Support Vector Machine (SVM)
  • NLP (LLM)
  • Model Evaluation, Hyperparameter Tuning

Final Year Project : Application of YOLO for detection and text extraction of HSRP/ NON HSRP number plates in video using Google Colab (Final year Project 2021)

In this project an attempt was made to identify and segregate High Security Registration Plates (HSRP) and non HSRP plates of vehicles in video and to extract the license plate number to store in a database for government surveillance and compliance purposes using YOLOv3 for Object Detection. I have successfully customized the YOLOv3 detection module to suit our case. The model can accurately detect HSRP/ non HSRP plates in video at 30 fps. A precision level of 90% was achieved. My model encloses the HSRP hologram in a bounding box within the number plate bounding box for most HSRP plates. The second objective of this project was to extract the number plate text. I have used EasyOCR to successfully extract the number plate text. Also, a strategy was incorporated in the program to reduce the number of frames to be processed. This has significantly increased the efficiency of the detector system. In this project as we are just detecting the number plates, processing just one frame per second suffices to capture all the vehicles rather than processing all the 30 frames per second as in the original video. This process not only reduces the file sizes of the output video, but also reduces the folder size of the XML detections thus reducing computational efforts significantly and increases detection speed. This entire project has been executed on Google Colab, which is a free Jupyter notebook environment that runs in the cloud and stores its notebooks on Google Drive., 2021

Certification

  • Machine Learning, Big Data Modelling & Management, Big Data Integration & Processing, Graph Analytics courses from Coursera (2020-2021)

Languages

Hindi
First Language
English
Proficient (C2)
C2

Personal Information

Date of Birth: 07/03/1999

Websites

Timeline

Graduate Engineer- Digital (AI/ML)

Mott MacDonald India
10.2021 - Current

Certification - Machine Learning And Deep Learning

IIT Delhi

B.Tech - Computer Science And Engineering

Manipal Institute of Technology, MAHE

Higher Secondary -

Delhi Public School Dwarka

Similar Profiles

CREATE PROFILE
NRIP GUPTA