Research

Conducted multi-modal place recognition by fusing 3D point clouds geometric features with image features from Vision-Language Models (VLMs) using late fusion and bilinear pooling to improve robustness and average recall.  

Chien Erh Lin, Minghan Zhu, Maani Ghaffari

RA-L 2024 [Paper] [GitHub]

We propose SE3ET, an SE(3)-equivariant registration framework that employs equivariant point convolution and equivariant transformer design to learn expressive and robust geometric features. 

Chien Erh Lin, Jingwei Song, Ray Zhang, Minghan Zhu, Maani Ghaffari

CoRL 2022 [Paper] [Github]

Learned SE(3)-invariant global descriptors are more robust to matching inaccuracy and failure in severe rotation and translation configurations.

Ray Zhang, Tzu-Yuan Lin, Chien Erh Lin, Steven A. Parkison, William Clark, Jessy W. Grizzle, Ryan M. Eustice, Maani Ghaffari

ICRA 2021 [Paper][arXiv][GitHub][Video]

A novel nonparametric rigid point cloud registration framework that jointly integrates geometric and semantic measurements such as color or semantic labels into the alignment process and does not require explicit data association.

Chien Erh Lin, Tzu-Yuan Lin, Yichebg Tao, Ray Zhang, Lu Gan, Maani Ghaffari

Added camera and computational resources to perform semantic mapping on the MIT mini cheetah.

Chien Erh Lin, Maani Ghaffari

Used SIFT-flow as additional labels to improve Continuous Visual Odometry (CVO).

[GitHub]

V. Swetha E. Jeganathan*, Chien Erh Lin*, Hojun Son, Divya S. Krishnagiri, Yumou Wei, James D. Weiland

EMBC 2022 [Paper] [GitHub]

Developed a computer vision-based system that can navigate someone visually impaired to the targeted door with object detection, object tracking, and visual SLAM.

Chien Erh Lin, Eduardo Iscar Rüland, Matthew Johnson-Roberson

Reconstructed an underwater scene in 3D, using structured light systems that project light patterns on the objects and enable the reconstruction of its structure, despite the low texture and low contrast of underwater images.

Dan Jeric Arcega Rustia, Chien Erh Lin, Jui-Yung Chung, Yi-Ji Zhuang, Ju-Chun Hsu, Ta-Te Lin

Journal of Asia-Pacific Entomology 2020 [Paper]

Constructed an IoT-based wireless imaging and sensor node system for remote greenhouse pest monitoring, with 90% accuracy of pest classification using deep neural network, enabling farmers to notice the amount of pest change in real-time easily.

Chien Erh Lin, Ta-Te Lin

Built a Plant Growth Monitoring System to monitor the growth of Cordyceps militaris using Qt, OpenCV and Raspberry Pi.

 Hao-Ming Hsiao, Tzu-Yuan Lin, Chien-Erh Lin, Han-Yu Lee, Yi-Ping Wang

Applied Sciences 2017 [Paper] 

A novel nitinol “spherical occlusion device” concept that can reduce the blood flow to cancer cells by deploying it in the upstream hepatic artery supplying blood to the liver. Moreover, it could carry multiple chemotherapy or radioactive drugs for delivery directly to the target site.