Dingding Cai

I am a PhD candidate at Computer Vision Group, Tampere University, Finland, where I am working on 3D computer vision and machine learning under the supervision of Prof. Esa Rahtu and Prof. Janne Heikkilä.

I obtained my Master's degree in Data Engineering and Signal Processing at Tampere University of Technology supervised by Prof. Joni Kämäräinen and Dr. Ke Chen in 2017. Prior to my doctoral study, I was a computer vision algorithm engineer working in Misshfresh Limited, Shenzhen, China.

Email (dingding.cai@tuni.fi)  /  Google Scholar  /  CV

profile photo
Research

I'm interested in various 2D/3D computer vision tasks, in particular, 3D scene understanding, 6D object pose estimation and tracking.

GS-Pose: Generalizable Segmentation-based 6D Object Pose Estimation with 3D Gaussian Splatting
Dingding Cai, Janne Heikkilä, Esa Rahtu
3DV, 2025
Project page / Paper / Code

We propose a unified framework for locating and estimating the 6D pose of novel objects. The proposed apporach leverages object co-segmentation and 3D Gaussian splatting for RGB-based 6D object pose estimation.

MSDA: Monocular Self-supervised Domain Adaptation for 6D Object Pose Estimation
Dingding Cai, Janne Heikkilä, Esa Rahtu
SCIA, 2023
Paper

We propose a practical self-supervised domain adaptation approach that takes advantage of real RGB(-D) data without needing real pose labels.

SC6D: Symmetry-agnostic and Correspondence-free 6D Object Pose Estimation
Dingding Cai, Janne Heikkilä, Esa Rahtu
3DV, 2022
Paper / Code

We present an efficient symmetry-agnostic and correspondence-free framework, referred to as SC6D, for 6D object pose estimation from a single monocular RGB image. SC6D requires neither the 3D CAD model of the object nor any prior knowledge of the symmetries.

OVE6D: Object Viewpoint Encoding for Depth-based 6D Object Pose Estimation
Dingding Cai, Janne Heikkilä, Esa Rahtu
CVPR, 2022
Project page / Paper / Code

We propose a universal object 6D pose estimation model, called OVE6D, purely trained on synthetic 3D objects from ShapeNet and generalizing remarkably well to unseen objects without needing any parameter optimization.

Convolutional Low-resolution Fine-grained Classification
Dingding Cai, Ke Chen, Yanlin Qian, Joni-Kristian Kämäräinen
Pattern Recognition Letters, 2019
Paper / Code

We propose a novel resolution-aware deep model which combines convolutional image super-resolution and convolutional fine-grained classification into a single model in an end-to-end manner.