Depth-Aware Networks for Multi-Organ Lesion Detection in Chest CT Scans

• Published in: Bioengineering (Basel) Vol. 11; no. 10; p. 998
• Main Authors: Zhang, Han, Chung, Albert C S
• Format: Journal Article
• Language: English
• Published: Switzerland MDPI AG 03-10-2024; MDPI
• Subjects: Back propagation; Chest; Computed tomography; computer-aided diagnosis; Computer-aided medical diagnosis; convolutional neural network; CT imaging; Datasets; Disease; Image processing; Image resolution; Lesions; Lung cancer; Lymphatic system; Machine learning; Medical imaging; Medical imaging equipment; Medical research; Medical screening; Metastasis; Methods; Molds; multi-organ lesion detection; Neural networks; China; Germany; computer-aided diagnosis; convolutional neural network; multi-organ lesion detection
• ISSN: 2306-5354; 2306-5354
• DOI: 10.3390/bioengineering11100998

Computer tomography (CT) scans' capabilities in detecting lesions have been increasing remarkably in the past decades. In this paper, we propose a multi-organ lesion detection (MOLD) approach to better address real-life chest-related clinical needs. MOLD is a challenging task, especially within a large, high resolution image volume, due to various types of background information interference and large differences in lesion sizes. Furthermore, the appearance similarity between lesions and other normal tissues demands more discriminative features. In order to overcome these challenges, we introduce depth-aware (DA) and skipped-layer hierarchical training (SHT) mechanisms with the novel Dense 3D context enhanced (Dense 3DCE) lesion detection model. The novel Dense 3DCE framework considers the shallow, medium, and deep-level features together comprehensively. In addition, equipped with our SHT scheme, the backpropagation process can now be supervised under precise control, while the DA scheme can effectively incorporate depth domain knowledge into the scheme. Extensive experiments have been carried out on a publicly available, widely used DeepLesion dataset, and the results prove the effectiveness of our DA-SHT Dense 3DCE network in the MOLD task.