Part of my work for my Bachelor's Thesis Project on Counterfactual Reasoning for Videos.

Towards Task Understanding in Visual Settings

A panoptic segmentation deep learning architecture for sim2real autonomous driving scene understanding

Investigating the utility of VR for spatial understanding in surgical planning: evaluation of head-mounted to desktop display

Pixel segmentation of roads from dashboard camera using Fully Convolutional Network

Label the pixels of a road in images using a Fully Convolutional Network (FCN).

This GitHub repository focuses on an integrated approach to scene classification and image caption generation, aiming to improve the accuracy of scene evaluation in computer vision applications.

[IAC 2023] This repository contains the code used in our paper, "AstrobeeCD: Change Detection in Microgravity with Free-Flying Robots." This method is useful for detecting 3D scene changes given a 3D model, a sequence of images, and a sequence of camera poses.

Implementation of inference for self-driving car perception tasks like object detection

Scene understanding is a vital aspect of safe and effective autonomous driving. And with the increase of high-quality datasets in recent years, the models have sufficient data to train on. However, the underlying models are important factors in determining the overall effect

Semantic Segmentation

This is the code for our ICCV'19 paper on cross-modal learning and retrieval.

Python implementation for scaled layout estimation from non-central panoramas

This repository contains code for image-to-image retrieval using different architectures. We test dilated CNN for the problem as well.

A Brief Tutorial on LiDAR data visualisation and classification

This study investigates the performance effect of using recurrent neural networks (RNNs) for semantic segmentation of urban scene images, to generate a semantic output map with refined edges. We proposed three deep neural network architectures using recurrent neural networks and evaluated them on the Cityscapes dataset. All three proposed archit…

Multi-Agent VQA: Exploring Multi-Agent Foundation Models on Zero-Shot Visual Question Answering

A visual scene dataset created based on the "VisualGenome" [https://visualgenome.org/] dataset.

Scripts, figures, and working notes for the participation in ImageCLEFmedical, part of the 14th CLEF Conference, 2023.