Quad-YOLOv5: Improved YOLOv5 for Liver Lesion Detection on Bio Medical CT Images
Keywords:
Quad-YOLOV5, Deep Learning, DeepLesion, Lesion Detection, CT Images, LocalizationAbstract
The liver is an important organ in the human body, performing numerous vital functions that are essential for overall health and well-being. Accurately identifying lesions in medical CT scans has long been one of the most difficult problems in the field of medical image analysis. Appropriate treatment and management strategies can be implemented to address the underlying liver condition and optimize patient outcomes. The proposed method for Liver Lesion Detection, utilizes the DeepLeison dataset which contains many biomedical CT scan images with a variety of liver pathologies. The method proposes Quad-YOLOv5 which is based on popular driven object detection deep learning model called YOLOV5 (You Only Look Once) model. We build up the medical image dataset of Liver Lesion by collecting 6335 CT images by augmentation. To enhance the performance of the Quad-YOLOv5 model, we have implemented data augmentation techniques and conducted extensive experiments using the DeepLeison dataset. Our findings demonstrate that the model exhibits strong performance coupled with remarkable interpretability. Through meticulous experimentation, we have refined the model's capabilities, ensuring that it delivers superior results in lesion detection tasks while maintaining a high level of interpretability. Our model is trained and evaluated based on its performance. The precision and recall for the model are 96% and 93%. It is obtained that Quad-YOLOv5 model is with Mean Average Precision (mAP50) of 97%. The model increases the efficiency and accuracy of diagnosing and treating liver lesions. It can be incorporated into existing clinical workflows to aid radiologists in the interpretation of CT scans.
Downloads
References
Ke Yan, Mohammadhadi Bagheri, Ronald M. Summers, "3D Context Enhanced Region-based Convolutional Neural Network for End-to-End Lesion Detection", MICCAI, 2018.
Youbao Tang, Adam P. Harrison, Mohammadhadi Bagheri, Jing Xiao, Ronald M. Summers,"Semi-Automatic RECIST Labeling on CT Scans with Cascaded Convolutional NeuralNetworks", MICCAI, 2018.
Jinzheng Cai*, Youbao Tang*, Le Lu, Adam P. Harrison, Ke Yan, Jing Xiao, Lin Yang, RonaldM. Summers, "Accurate Weakly-Supervised Deep Lesion Segmentation using Large-ScaleClinical Annotations: Slice-Propagated 3D Mask Generation from 2D RECIST", MICCAI,2018.
Amita Das, Priti Das, S. S. Pandaand Sukanta Sabut “Detection of Liver Cancer Using Modified Fuzzy Clustering and Decision Tree Classifier in CT Images” Pattern Recognition and Image Analysis, 2019, Vol. 29, No. 2, pp. 201–211. © Pleiades Publishing, Ltd., 2019.
Ke Yan, Yifan Peng, Veit Sandfort, Mohammadhadi Bagheri, Zhiyong Lu, and Ronald M.Summers, “Holistic and Comprehensive Annotation of Clinically Significant Findings onDiverse CT Images: Learning from Radiology Reports and Label Ontology,” CVPR, 2019.
Zhiqi Bai, Huiyan Jiang, Siqi Li, And Yu-Dong Yao, “Liver Tumor Segmentation Based On Multi-Scale Candidate Generation And Fractal Residual Network” ,June, 2019, Volume 7, 2019, Ieee Transaction.
Xin Dong, Yizhao Zhou, Lantian Wang, Jingfeng Peng,Yanbo Lou,And Yiqun Fan “Liver Cancer Detection Using Hybridized Fully Convolutional Neural Network Based On Deep Learning Framework” Ieee Access ,Special Section On Deep Learning Algorithms For Internet Of Medical Things, July 24, 2020, Natural Science Foundation of Zhejiang Province under Grant LY17H160024.
M. Chung, J. Lee, M. Lee, J. Lee, and Y.-G. Shin, ‘‘Deeply self-supervised contour embedded neural network applied to liver segmentation,’’ Comput. Methods Programs Biomed., vol. 192, Aug. 2020, Art. no. 105447
Downloads
Published
How to Cite
Issue
Section
License
![Creative Commons License](http://i.creativecommons.org/l/by-sa/4.0/88x31.png)
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
All papers should be submitted electronically. All submitted manuscripts must be original work that is not under submission at another journal or under consideration for publication in another form, such as a monograph or chapter of a book. Authors of submitted papers are obligated not to submit their paper for publication elsewhere until an editorial decision is rendered on their submission. Further, authors of accepted papers are prohibited from publishing the results in other publications that appear before the paper is published in the Journal unless they receive approval for doing so from the Editor-In-Chief.
IJISAE open access articles are licensed under a Creative Commons Attribution-ShareAlike 4.0 International License. This license lets the audience to give appropriate credit, provide a link to the license, and indicate if changes were made and if they remix, transform, or build upon the material, they must distribute contributions under the same license as the original.