Browsing by Subject "Lung segmentation"

Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    70329 Automated Lungs Segmentation and Airways Skeletonization from CT Scans in Patients with Cystic Fibrosis
    (Cambridge University Press, 2021) Chie, Juan A. Chong; Territo, Paul R.; Salama, Paul; Medicine, School of Medicine
    ABSTRACT IMPACT: Improve healthcare of patients with Cystic Fibrosis by reducing the time needed to generate results. OBJECTIVES/GOALS: We developed an automated framework capable of segmenting the lungs, extract the airways, and create a skeletonize map of the airways from CT scans of Cystic Fibrosis patients. As future expansion, the framework will be expanded to measure the airways diameters, detect the abnormal airways, and count the number of visible airways generations. METHODS/STUDY POPULATION: For this study, 35 CT scans from CF patients with different levels of severity were used to test the developed framework. The lungs segmentation was performed using an algorithm based on Gaussian Mixture Models for mild cases, and for severe cases a technique that uses convex hull and the recurrent addition of ‘dots’ was implemented. The airways extraction was performed using a 26-points connected components algorithm in conjunction with a curve fitting technique over the histogram of voxel values. Medial axis transform was used to perform the skeletonization of the extracted airways, and airways diameters determined via ray-casting. RESULTS/ANTICIPATED RESULTS: The framework was able to correctly obtain the segmented lungs in all 35 sample volumes regardless of disease severity. In contrast, it tends to fail to skeletonize the airways for severe cases where the framework is unable to differentiate between abnormal lungs conditions and dilated airways. Fine tuning is required to achieve better results. The expected result of the future implemented sections of the framework are focused to characterize the extracted airways by: 1) measuring the airways diameters; 2) detect and count the number of abnormal airways sizes; and 3) count the number of visible airways branching which will permit determination of stage and grade of the lungs of CF patients. DISCUSSION/SIGNIFICANCE OF FINDINGS: The proposed framework allows a fast and reproducible way to segment the lungs and create a skeletonized map of the airways that are independent of clinical training. In addition, this framework will be extended to obtain measurements of airway dilation and branching level, which could provide a deeper insight of the airways in CF patients.
  • Thumbnail Image
    Item
    A Novel Method for 3D Lung Tumor Reconstruction Using Generative Models
    (MDPI, 2024-11-20) Najafi, Hamidreza; Savoji, Kimia; Mirzaeibonehkhater, Marzieh; Moravvej, Seyed Vahid; Alizadehsani, Roohallah; Pedrammehr, Siamak; Electrical and Computer Engineering, Purdue School of Engineering and Technology
    Background: Lung cancer remains a significant health concern, and the effectiveness of early detection significantly enhances patient survival rates. Identifying lung tumors with high precision is a challenge due to the complex nature of tumor structures and the surrounding lung tissues. Methods: To address these hurdles, this paper presents an innovative three-step approach that leverages Generative Adversarial Networks (GAN), Long Short-Term Memory (LSTM), and VGG16 algorithms for the accurate reconstruction of three-dimensional (3D) lung tumor images. The first challenge we address is the accurate segmentation of lung tissues from CT images, a task complicated by the overwhelming presence of non-lung pixels, which can lead to classifier imbalance. Our solution employs a GAN model trained with a reinforcement learning (RL)-based algorithm to mitigate this imbalance and enhance segmentation accuracy. The second challenge involves precisely detecting tumors within the segmented lung regions. We introduce a second GAN model with a novel loss function that significantly improves tumor detection accuracy. Following successful segmentation and tumor detection, the VGG16 algorithm is utilized for feature extraction, preparing the data for the final 3D reconstruction. These features are then processed through an LSTM network and converted into a format suitable for the reconstructive GAN. This GAN, equipped with dilated convolution layers in its discriminator, captures extensive contextual information, enabling the accurate reconstruction of the tumor's 3D structure. Results: The effectiveness of our method is demonstrated through rigorous evaluation against established techniques using the LIDC-IDRI dataset and standard performance metrics, showcasing its superior performance and potential for enhancing early lung cancer detection. Conclusions: This study highlights the benefits of combining GANs, LSTM, and VGG16 into a unified framework. This approach significantly improves the accuracy of detecting and reconstructing lung tumors, promising to enhance diagnostic methods and patient results in lung cancer treatment.