Publication

Needle Localization and Segmentation for Radiofrequency Ablation of Liver Tumors under CT image Guidance

Le Quoc Anh, Luu Manh Ha, Theo van Walsum, Adriaan Moelker, Dao Viet Hang, Pham Cam Phuong and Vu Duy Thanh

2022 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC). IEEE.

Radiofrequency ablation (RFA) of liver cancer under computer tomography (CT) guidance is a minimally invasive procedure in which CT images are utilized to guide the physician in introducing the needle into the target lesion. However, the adequate visualization of the needle and anatomy is hampered by the 2D slide based-view used in the current clinical practice. Thus, due to the lack of 3D information, the physician requires high experience and more interaction with the guidance systems to envision the needle’s position in the liver, which is inconvenient in clinical practice. In this study, we propose a method for robust needle segmentation using CT images to improve the visualization of the needle during the intervention. The method utilizes a convolutional neural network (CNN) to detect the needle in orthogonal 2D projections of the CT image to construct the needle volume of interest (VOI). Subsequently, a patch-based 3D CNN is applied to segment the needle. We evaluate the method’s accuracy using Dice score (DSC ), Hausdorff distance (HD), the needle shaft error E_{shaft}, and needle tip error E_{tip}. The results show that the proposed method achieves the means of DSC, HD, E_{tip}, E_{shaft} and processing time of 0.89, 3.3 mm, 0.9 mm, 0.43 mm, and 2.6 seconds, respectively. We conclude that the proposed method is feasible for improving needle visualization in the interventional room

Efficient Type and Polarity Classification of Chromosome Images using CNNs: a Primary Evaluation on Multiple Datasets

Le Quoc Anh, Vu Duy Thanh, Nguyen Huu Hoang Son, Doan Thi Kim Phuong, Luong Thi Lan Anh, Do Thi Ram, Nguyen Thanh Binh Minh, Tran Hoang Tung, Nguyen Hong Thinh, Le Vu Ha, Luu Manh Ha

2022 IEEE Ninth International Conference on Communications and Electronics (ICCE)

Abstract | Paper

Karyotyping is critical for screening genetic diseases in an early stage. However, the manual karyotyping process is a labor-intensive task. This paper focuses on automatic chromosome image classification, which is a step in karyotyping. We propose a convolutional neural network architecture for efficient type and polarity classification of the chromosome image, namely ETPC, which is leveraged from the EfficientNet family’s development. The ETPC’s classifier with a weighted classification loss function are designed for efficiency in the training process. We perform our experiment with two training scenarios on four clinical datasets. The experiment on a dataset of 28,225 original chromosome images demonstrates that the proposed network achieved comparable results, with an accuracy of 95.3% for the type classification and 99% for the polarity classification, while having a significantly smaller number of parameters than state-of-the-art methods. Furthermore, the experiment on multiple datasets shows that the proposed network can dramatically improve the classification accuracy on new datasets with a small amount of fine-tuning data.