Obafemi Jinadu

Obafemi Jinadu

PhD candidate @ Tufts University.

Hello there, I’m Obafemi image

I’m a 4th year PhD candidate at Karen Panetta’s Vision Sensing and Simulation Lab at Tufts University, where I work on computer vision, multimodal machine learning and generative AI to address complex, real-world problems.

Research

My research is centered on developing cutting-edge machine learning, and deep learning techniques for vision and multimodal understanding (vision-language), toward deploying reliable, robust, and trustworthy models that achieve state-of-art performance.

image Attention-based two-stream high-resolution networks for building damage assessment from satellite imagery.
Victor Oludare, Landry Kezebou, Obafemi Jinadu (corresponding author), Karen Panetta, Sos Agaian: In Multimodal Image Exploitation and Learning 2022 (Vol. 12100, pp. 224-239). SPIE, 2022.
Poster Presentation

image Instant-level vehicle speed and traffic density estimation using deep neural network.
Obafemi Jinadu (corresponding author), Victor Oludare, Srijith Rajeev, Landry Kezebou, Karen Panetta, Sos Agaian: In Multimodal Image Exploitation and Learning 2023, vol. 12526, pp. 125-138. SPIE, 2023.
Poster Presentation. Project GitHub repository can be found here

image An Impact Study of Deep Learning-based Low-light Image Enhancement in Intelligent Transportation Systems.
Obafemi Jinadu (corresponding author), Srijith Rajeev, Karen Panetta, Sos Agaian: In Multimodal Image Exploitation and Learning 2024 (Vol. 13033, pp. 154-171). SPIE. Project GitHub repository can be found here

image An Individualized Machine Learning Approach for Human Body Weight Estimation Using Smart Shoe Insoles.
Foram Sanghavi, Obafemi Jinadu, Victor Oludare, Karen Panetta, Landry Kezebou, Susan B. Roberts: Sensors 2023, 23(17) Special Issue Wearables and Artificial Intelligence in Health Monitoring.

Personal & Course-based Machine Learning Adjacent Projects

  • My DDPM re-implementation (Github Repository here).
    I re-implemented the paper denoising diffusion probabilistic model (DDPM) from scratch. The repository explores the strengths and weaknesses of DDPMs along with some ablation studies.
  • VisionTransformer (ViT) Re-implementation from Scratch (Github Repository here).
    I re-implemented the paper An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalefrom scratch. The repository explores the Vision Transformer Architecture in detail, mapping the theory into code and thus cementing my understanding of the components that make up this architecture.

  • Probabilistic Model for Amur Tiger Re-identification my paper | course website
    I developed an animal re-identification deep-learning model that classifies 107 amur tiger entities using a Maximum A Posteriori (MAP) estimation model to counter the overfitting challenge of traditional maximum likelihood estimation-based models with severely limited and imbalanced training data.
    Instructor comment: “very nice overall! Would be excited to talk more if you want to follow up on this project. I think could be a great potential future CP for this class, among other things (maybe also a cool research paper)” ~ Michael C. Hughes
    Statistical Pattern Recognition, Spring 2023 Score 104/100 Grade A

  • Unified Self-Supervised Algorithm for Speech, Vision, and Text: Meta’s data2vec Project paper re-write here. Adapted from original paper
    Machine Learning, Spring 2022 Grade A+
  • Natural Language Processing in Classifying Review Sentiment with Bag-of-Words Features project documentation
    Machine Learning, Spring 2022 Grade A+

Recent News & Updates