Research
I'm interested in computer vision, Self-supervised learning, generative models, and their application to a range of problems. Currently I'm focusing on 3D human pose/mesh estimation from monocular videos. Some of my works are mentioned below.
|
|
STARS: Self-supervised 3D Action Recognition with Contrastive Tuning
Soroush Mehraban,
Mohammad Javad Rajabi,
Babak Taati
arxiv
project page
/
arXiv
STARS enhances the Mask Autoencoder (MAE) approach in self-supervised learning by applying contrastive tuning. We also show that MAE approaches fail in few-shot settings and achieve improved performance by using the proposed method.
|
|
Benchmarking Skeleton-based Motion Encoder Models for Clinical Applications: Estimating Parkinson's Disease Severity in Walking Sequences
Vida Adeli,
Soroush Mehraban,
Irene Ballester,
Yasamin Zarghami,
Andrea Sabo,
Andrea Iaboni,
Babak Taati
FG, 2024
Code
/
arXiv
Evaluating recent motion encoders for the task of parkinsonism severity estimation (UPDRS III gait)
|
|
MotionAGFormer: Enhancing 3D Pose Estimation with a Transformer-GCNFormer Network
Soroush Mehraban
Vida Adeli,
Babak Taati
WACV, 2024
Code
/
video
/
arXiv
Estimating 3D locations of 17 main joints from a monocular video.
|
|