Yasiru Ranasinghe

I am a 4th year Ph.D. student at the Vision & Image Understanding Lab, Johns Hopkins University, advised by Dr. Vishal M. Patel. My research focuses on scene understanding, object detection, self-supervised learning, and leveraging foundation models for agent-based AI. I am currently spending the summer as a research intern at Apple.

Email  /  Google Scholar  /  LinkedIn  /  Download CV

Profile Photo

Research

My research focuses on advancing visual scene understanding through object detection, self-supervised learning, and the use of foundation models. I am particularly interested in building vision systems that generalize across tasks and environments, with an emphasis on integrating foundation models into agent-based AI for perception.


News

  • 2025
    • May: Started summer internship at Apple.
    • May: One paper accepted at AVSS'25.
    • February: One paper accepted at CVPR'25.
  • 2024
    • March: One paper accepted at IEEE FG'24.
    • February: Two papers accepted at CVPR'24.

Publication List

CVPR

SINR main figure
SINR pipeline

SINR: Sparsity Driven Compressed Implicit Neural Representations

Authors: Dhananjaya Jayasundara, Sudarshan Rajagopalan, Yasiru Ranasinghe, Trac D. Tran, and Vishal M. Patel

CVPR, 2025  

This work presents a novel compression approach for implicit neural representations (INRs) by encoding their weight space using a sparse high-dimensional dictionary. Unlike prior INR compression methods, SINR avoids transmitting learned dictionaries and remains compatible with existing INR frameworks. It achieves significant storage savings while preserving reconstruction quality across diverse signal types including images, occupancy fields, and NeRFs.

CrowdDiff: Multi-hypothesis Crowd Density Estimation using Diffusion Models

Authors : Yasiru Ranasinghe, Nithin Gopalakrishnan Nair, Wele Gedara Chaminda Bandara, and Vishal M. Patel

CVPR, 2024  

The paper proposes a novel approach to perform crowd counting with multi-hypothesis aggregation using denoising diffusion probabilistic models. The approach outperforms existing state-of-the-art methods for crowd counting.

MonoDiff: Monocular 3D Object Detection and Pose Estimation with Diffusion Models

Authors : Yasiru Ranasinghe, Deepti Hegde, and Vishal M. Patel

CVPR, 2024  

The paper proposes using diffusion models to perform monocular 3D object detection and pose estimation. MonoDiff does not require additional modalities to generate intermediate representations to produce box parameters.

Preprints

Conferences

Journals


Template taken from here. Last updated October 2023.