Xingxing Liu

Xingxing Liu

Shing-shing Lyoo

I am a final-yeal Ph.D. candidate. My research mainly focus on AI Healthcare.

Publications

You can also find my articles on my Google Scholar profile.

Journal Articles

Multi-View Fusion-Based 3D Object Detection for Robot Indoor Scene Perception

Published in Sensors, 2021

Accurate 3D object detection enables service robots to have 3D scene perception in cluttered indoor environments, but it’s usually a challenging task. In this paper, we proposed a two-stage 3D object detection algorithm which is to fuse multiple views of 3D object point clouds in the first stage and to eliminate unreasonable and intersection detections in the second stage. For each view, 3D object bounding box estimation has four steps: (1) 2D object semantic segmentation and 3D object point clouds reconstruction; (2) using Locally Convex Connected Patches (LCCP) method to segment the object from the background; (3) calculating the main object orientation with Manhattan Frame estimation method; (4) constructing 3D object bounding box. An object database is created and refined as more multi-view point clouds of the same object are fused. Incorrect and intersecting objects are removed from the object database based on prior knowledge. Experiments performed on both SceneNN dataset and a real indoor environment show the high accuracy and stability of our proposed method.

Recommended citation: Wang, L.; Li, R.; Sun, J.; Liu, X.; Zhao, L.; Seah, H.S.; Quah, C.K.; Tandianus, B. Multi-View Fusion-Based 3D Object Detection for Robot Indoor Scene Perception. Sensors 2019, 19, 4092. https://doi.org/10.3390/s19194092
Download Paper

Conference Papers

Deep Convolutional Feature-Based Fluorescence-to-Color Image Registration

Published in 2021 IEEE International Symposium on Medical Measurements and Applications (MeMeA), 2021

Fluorescence guided surgery (FGS) combines functional information (fluorescence imaging) and structural information (color imaging) to improve surgery performance. Fluorescence-to-color image registration plays fundamental role in FGS. In this paper, we used VGG16 to extract image features from color and fluorescence images, and feature descriptors were built with these features. Then, keypoint matching was conducted to build correspondence between color image and fluorescence image. Finally, fluorescence-to-color image registration was achieved based on the matched keypoint pairs. Experimental results show that our method outperforms conventional feature based image registration algorithms, like SIFT, BRISK, SUFT, ORB.

Recommended citation: X. Liu, T. Quang, W. Deng and Y. Liu, "Deep Convolutional Feature-Based Fluorescence-to-Color Image Registration," 2021 IEEE International Symposium on Medical Measurements and Applications (MeMeA), Lausanne, Switzerland, 2021, pp. 1-6, doi: 10.1109/MeMeA52024.2021.9478607.
Download Paper | Download Slides

Application of Hybrid Network of UNet and Feature Pyramid Network in Spine Segmentation

Published in 2021 IEEE International Symposium on Medical Measurements and Applications (MeMeA), 2021

Spine segmentation plays an important role in spinal disease diagnosis and treatment, also, it is a fundamental procedure in some spine surgical navigation systems. In this paper, we proposed a hybrid network of Feature Pyramid Network (FPN) and UNet, and used it for vertebral body segmentation. Experiments were conducted with a T2-weighted lower spine MRI dataset. Experimental results show that our proposed network outperforms UNet and several other UNet based networks in spine segmentation. Quantitative analysis shows that segmentation accuracy of 99.5% can be achieved with this network.

Recommended citation: X. Liu, W. Deng and Y. Liu, "Application of Hybrid Network of UNet and Feature Pyramid Network in Spine Segmentation," 2021 IEEE International Symposium on Medical Measurements and Applications (MeMeA), Lausanne, Switzerland, 2021, pp. 1-6, doi: 10.1109/MeMeA52024.2021.9478765.
Download Paper | Download Slides