Wele Gedara Chaminda Bandara
I am a Computer Vision Scientist (Engineer) at Apple, where I develop solutions to computer vision problems in the next generation of Apple devices.
I completed my Ph.D. in Electrical and Computer Engineering from Johns Hopkins University, advised by Prof. Vishal M. Patel. My dissertation research focused on spatiotemporal representation learning techniques for videos and remote sensing applications. Additionally, I worked on projects involving Generative AI, particularly with diffusion models, and low-level vision such as image segmentation and image super-resolution. During my Ph.D. studies, I had the opportunity to participate in two amazing summer internships at GetZippin in 2022 and Apple in 2023.
Updates ...
March 5, 2024: Our paper Attention Prompt Tuning: Parameter-efficient Adaptation of Pre-trained Models for Spatiotemporal Modeling has been accepted at FG'24, Istanbul, Turkey!
February 9, 2024: I successfully defended my PhD, titled Efficient Spatiotemporal Representation Learning Techniques for Videos and Remote Sensing Data. Presentation slides available here.
Feb 27, 2023: Two papers accepted for the CVPR'23.
AdaMAE: Adaptive Masking for Efficient Spatiotemporal Learning with Masked Autoencoders (click here)
Unite and Conquer: Cross Dataset Multimodal Synthesis using Diffusion Models (click here)
February 26, 2024: Our paper CrowdDiff: Multi-hypothesis Crowd Density Estimation using Diffusion Models has been accepted at CVPR'24, Seattle, USA!
June 18-23, 2023: Presented our papers ("AdaMAE" and "Unite and Conquer") at CVPR'23 in Vancouver, Canada!
May 4, 2022: Our paper Orientation-guided Graph Convolutional Network for Bone Surface Segmentation has been accepted at MICCAI'22 (early acceptance).
February 19, 2024: Re-joined Apple as an AI/ML Engineer!
May 1, 2023: I joined Apple as video research intern during summer 2023!
April 29, 2022: Our paper SPIN Road Mapper: Extracting Roads from Aerial Images via Spatial and Interaction Space Graph Reasoning for Autonomous Driving has been selected for the finalists (top 3) of the ICRA 2022 Outstanding Automation paper!
March 1, 2022: Our paper HyperTransformer: A Textural and Spectral Feature Fusion Transformer for Pansharpening has been accepted at CVPR'22. Click here to find out more about this work.
Contact
Personal email: w.chamindabandara [at] gmail.com