Projects
1. Transport for London: Identifying physical assets on the London Underground from the point cloud.
In Collaboration with Transport for London and Alan Turing Institute, UK
Started Project Date: June 2024
This DSG challenge of identification of assets is the collaboration of industry “Transport for London (TfL)” with Alan Turing Institute. Transport for London (TfL) allocates £60-80 million annually for maintenance and renewal planning, which necessitates detailed knowledge of current track assets. This project aims to use Image Data (converted from point cloud data generated from track trolley scans) to detect and classify various track features, including:
-
Installed sleeper and rail types
-
Sleeper spacing
-
Location of track-side equipment (lubricators, point machines, signaling equipment)
-
Location of drainage systems
-
Accurate positioning of features along the rail (expansion switches, joints, welds)
-
Conductor rail (3rd and 4th rail) type, position, and endpoints.
Results: Proposed approach modified SAM-2 and U-Net image segmentation models while incorporating depth information, achieving over 95% precision and ~100% instance-level recall in identifying various assets of Underground Railways.
2. Segmentation and Classification of Mycetoma
Leading team of PhD Students and Postdoctoral Researchers
University of Leeds, UK.
Started Project Date: July 2024
Task 1: Detection of Mycetoma Grains
Task 2: Classification of Mycetoma Type
-
Actinomycetoma (caused by bacteria)
-
Eumycetoma (caused by fungi)
Conference Challenge
3. Skin Cancer Detection on Low Resolution Images
Collaboration with Generative AI @Google
Started Project Date: August 2024
Approach and Results will appear after publicly available proceedings
5. Team Science Project on Multiple Long Term Conditions (MLTC)
In Collaboration with King's College London, University of Edinburgh, Imperial College London, and University of Leeds, UK.
Started Project Date: Sept 2023
The term ‘multiple long-term conditions’ (MLTC) refers to the co-existence of two or more chronic conditions (physical or mental) in a person.
Here, my responsibilities includes:
-
Management of the project
-
Lead Machine Learning Research
Grant: NIHR Teams Science Grant (Accepted: 100K) to conduct Research on Multiple Long Term Conditions.
6. Knowledge Graph Based CPRD Project
In Collaboration with University of Edinburgh, and University of Leeds, UK.
Started Project Date: Sept 2023
This includes investigation of Convlutional and Graph Network based Approached for Electronic Health Records.
7. Network Architecture Search (NAS)
Associated with mindtrace(dot)ai
Duration: Jan to March 2023
Worked on: Network Architecture Search (NAS) using hyperparameter and model parameters for 3D pointcloud models (SPVCNN, PointNet).
Tools and Libraries: Pytorch, Tensorflow, Pytorch-Lightning and MLflow
8. IdahoPower POC
Associated with mindtrace(dot)ai
Duration: Nov 2022 to Feb 2023
Worked on: Worked on:
- SPVCNN
- Hierarchical Subclassification
- Tuning and other Optimization techniques
- Data Analytics (using Unsupervised approaches)
9. Netflix Movie Recommendation
Associated with Massachusetts Institute of Technology MIT
Date: Dec 2022 Link of Portfolio
Title: Netflix Movie Recommendation
Course: Recommendation Systems
The objective of this project is to build a recommendation system to recommend movies to users based on the ratings given to different movies by the users.
Skills and Tools: Collaborative filtering, Matrix factorization, Recommendation systems
10. Hotel Booking Cancellation Prediction
Associated with Massachusetts Institute of Technology MIT
Date: Nov 2022 Link of Portfolio
Project: Hotel Booking Cancellation Prediction
Course: Classification and Hypothesis Testing
The objective of the project is to analyse the data provided to find which factors have a high influence on booking cancellations, build a predictive model that can predict which booking is going to be cancelled in advance, and help in formulating profitable policies for cancellations and refunds.
Skills and Tools: Exploratory Data Analysis, Classification, Logistic Regression, Support vector Machines, Decision Trees, Random Forest
11. Pima Diabetes Analysis
Associated with Massachusetts Institute of Technology MIT
Date: Oct 2022 Link of Portfolio
Project: Pima Diabetes Analysis
Course: Foundations of Data Science
To analyse different aspects of Diabetes in the Pima Indians tribe.
Skills and Tools: Descriptive Statistics, Data Visualization, EDA
12. Semi-Supervised learning for 3D PointCloud
Associated with mindtrace(dot)ai
Duration: May to Oct 2022
Worked on:
- 3D Pointcloud Lidar Data
- Mean teacher (Semi Supervised Model)
- SPVCNN (3D CNN pointcloud model)
- PointNet (3D CNN pointcloud model)
- Batch Sampler
Tools and Libraries:
- Pytorch,
- Tensorflow,
- Pytorch-Lightning,
- and MLflow
13. esmart, GKN, Gestamp
Associated with mindtrace(dot)ai
Duration: Feb to May 2022
Worked on:
- Few Shot Detection (FsDet)
- Few Shot Segmentation (RePRI)
Tools and Libraries: Pytorch, Tensorflow, Pytorch-Lightning,
and MLflow
14. Defect Detection
Associated with mindtrace(dot)ai
Duration: Nov 2021 to Feb 2022
Worked on:
- GDXray dataset https://www.v7labs.com/open-datasets/gdxray
- Supervised Learning (Mask R-CNN)
- YOLOv5
- Google Cut Paste Method
15. Geodigital Anomaly Detection
Associated with mindtrace(dot)ai
Duration: June to Dec 2021
Worked on:
- Dino Facebook Research (Attention Models and Transformers) https://github.com/facebookresearch/dino
- Unsupervised Learning (PCA, DBScan, tsne)
- SCAN (Learning to Classify Images without Labels) https://arxiv.org/abs/2005.12320
16. CVPR 2021 RetailVision AliProducts Challenge
March-June 2021
Associated with mindtrace(dot)ai
Team placed Rank 11th.
Worked on:
- Metric Learning
- Cleanlab
- Aliproducts dataset (50000 classes)
In this we used: ResNext101, ResNet152, EfficientNet_B6_AP, HRNet_32w, Res2Net101 and Eca_NFNet_l0
17. Detecting Seen/Unseen Concepts Online while Reducing Response Time with/without Bounding Boxes using Domain Adaptive Multimedia Event Processing
PhD Research Scholar
April 2016 - March 2021
Thesis Title: Detecting Seen/Unseen Concepts Online while Reducing Response Time with/without Bounding Boxes using Domain Adaptive Multimedia Event Processing
Summary: Deep neural network-based techniques are effective for image classification, but the limitation of having to train classifiers for unseen concepts may increase the overall response-time for multimedia-based event processing models. This work focuses on foundational aspects of the problem of reducing response-time for online adaptive classifiers-based multimedia event processing which includes introducing object detection operators, standardization of the concept of response-time, identification, and proposed multiple IoMT based deep neural network models while using object detection specifically You Only Look Once (YOLO), Single Shot MultiBox Detector (SSD), and RetinaNet, and applying transfer learning. Lastly, I report the best possible performance of current object detection models for the online construction of classifiers. The major challenge in training deep neural network-based models is the need to collect many images with bounding box annotations, which is impossible for millions of unseen concepts. My final specific work is the design of first and fast detector for the training of unseen classes using only image-level labels with no bounding box annotations. It takes 10 min only to train an object detector.
Languages: Python, C, Shell Scripting (Linux Platform)
Libraries: TensforFlow, CUDA, cuDNN, Keras, OpenCV
Hardware Used: Nvidia Titan Xp GPU
Publication Outcomes: 3 Journals, 4 Conference papers (in CVPR 2022 selected for Oral Presentation, Journals of IEEE Access, Elsevier, and Springer; and other conferences including ICMR and DEBS)
-
Fellowship: Awarded Fellowship by Science Foundation Ireland Fellowship for PhD
-
Grant: NVIDIA GPU Grant for Titan Xp GPU by the Nvidia Corporation 2018
18. Image Segmentation Using Fuzzy Multi Criteria Decision Making
Masters Dissertation
July 2013 - June 2015
Title: Image Segmentation using Fuzzy Multi-Criteria Decision Making
Summary: Image segmentation refers to the separation of objects from the background. Practically it is impossible to design a segmentation algorithm that has 100% accuracy. In this dissertation, two methods of segmentation are proposed: the first one is the Improved Sobel Edge Detection algorithm and the second is the Falling Ball algorithm. Our Falling ball algorithm which is a region-based segmentation algorithm, an alternative to watershed transform (based on waterfall model) and applies Fuzzy Logic for the segmentation. Simulation results show that the proposed algorithms give superior performance over conventional Sobel edge detection methods and watershed segmentation algorithm.
Languages: C, Java, Shell Scripting (Linux Platform)
Publication Outcomes: 1 Journal with +150 citations
-
Fellowship from Graduate Aptitude Test in Engineering (GATE), All India Rank 597 out of 115,425 candidates.
19. Edge Detection using Ant Colony Optimization
Masters Project (6 months)
May 2014 - June 2015
Title: Edge Detection using Ant Colony Optimization
Summary: In this work, a multi-threading-based implementation of Ant Colony Optimization (ACO) is proposed for identifying edges in images. It combines multi-threading with ACO for increasing the randomness among the artificial ants. Simulation results show that the proposed method has significantly lower execution time as compared to conventional ACO for edge detection.
Languages: C, Java, Shell Scripting (Linux Platform)
Libraries: POSIX
Publication Outcomes: 1 Conference paper
20. Framework development and implementation of stereoscopic website
Bachelors Project (6 months)
Dec 2012 - May 2013
Title: Framework development and implementation of stereoscopic website
Summary: In this work, we worked on 3D images for the development of a Stereoscopic Website. We analyzed MPO and anaglyph 3D image formats. Moreover, we presented a new algorithm for obtaining depth information (for Depth-Map) pertaining to a depicted scene from a set of available pair of stereoscopic images.
Languages: C and HTML (Linux and Windows Platform)
Hardware Used: 3D television and stereoscopic glasses
Publication Outcomes: 1 arxiv paper and Book “Towards Stereoscopic Websites”
-
Prestigious Fellowship from IDB (till 4 years), Jeddah
-
Scholarship: Sir Syed Scholarship (till 4 years)
21. Simulation of M/G/1 Queue
Bachelors Project (6 months)
July 2011 - Dec 2011
Languages: C, Java, Shell Scripting
22. Text to Speech Converter
Bachelors Project (3 months)
Oct 2010 - Dec 2010
Languages: C and Java
4. DynAIRx: Artificial Intelligence (AI) Utilising advanced AI to support medicines optimisation in multimorbidity
In Collaboration with University of Manchester, University of Leeds, University of Liverpool, and University of Glasgow, UK.
Started Project Date: March 2023
DynAIRx (Artificial Intelligence for dynamic prescribing optimisation and care integration in multimorbidity) aims to develop new, easy to use, artificial intelligence (AI) tools that support General Practitioners (GPs) and pharmacists to find patients living with multimorbidity (two or more long-term health conditions) who might be offered a better combination of medicines.
Here, my responsibilities includes:
-
Leading Designing of Models for DynAIRx Project: Artificial Intelligence for Health Data Sciences @Institute of Health Data Sciences in collaboration with University of Manchester, University of Leeds, Alan Turing Institute, University of Liverpool, and University of Glasgow.
-
Leading Postdocs for Codelists Groupings of all Conditions
-
Other responsibilities: Co-Supervise PhD students, Teaching/Labs for Machine Learning, Statistics, Data Science etc.
Graph Neural Networks
Grant: National Institute for Health Research (NIHR) Grant, UK Research and Innovation (UKRI), worth £2.8 million.
Here I am working on Graph Neural Networks, ELectronics Health Records, Statistics, and Visualisations.