Paediatric Frontal Chest Radiograph Screening with Fine-Tuned Convolutional Neural Networks

Gerrand, Jonathan; Williams, Quentin; Lunga, Dalton; Pantanowitz, Adam; Madhi, Shabir А.; Mahomed, Nasreen

doi:10.1007/978-3-319-60964-5_74

Cited by 2 publications

(1 citation statement)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It is surprisingly concluded by many works that exploiting the transferibility of pre-trained CNN models can yield better results [29] with even shorter training time to fine-tune the adapted parameters for a different task or with different data [10]. In [30], [31], various medical imaging modalities were tested including training from scratch and as well as finetuning of pre-trained models that were trained with natural RGB images. The comprehensive comparisons have shown that finetuning pre-trained models, even with considerably different medical images, provides not only better image classification performance and segmentation tasks, but also help achieve training convergence faster.…”

Section: Pre-trained Modelsmentioning

confidence: 99%

Building Extraction at Scale using Convolutional Neural Network: Mapping of the United States

Yang¹,

Yuan²,

Lunga³

et al. 2018

Preprint

Self Cite

View full text Add to dashboard Cite

Establishing up-to-date large scale building maps is essential to understand urban dynamics, such as estimating population, urban planning and many other applications. Although many computer vision tasks has been successfully carried out with deep convolutional neural networks, there is a growing need to understand their large scale impact on building mapping with remote sensing imagery.Taking advantage of the scalability of CNNs and using only few areas with the abundance of building footprints, for the first time we conduct a comparative analysis of four stateof-the-art CNNs for extracting building footprints across the entire continental United States. The four CNN architectures namely: branch-out CNN, fully convolutional neural network (FCN), conditional random field as recurrent neural network (CRFasRNN), and SegNet, support semantic pixel-wise labeling and focus on capturing textural information at multi-scale. We use 1-meter resolution aerial images from National Agriculture Imagery Program (NAIP) as the test-bed, and compare the extraction results across the four methods. In addition, we propose to combine signed-distance labels with SegNet, the preferred CNN architecture identified by our extensive evaluations, to advance building extraction results to instance level. We further demonstrate the usefulness of fusing additional near IR information into the building extraction framework. Large scale experimental evaluations are conducted and reported using metrics that include: precision, recall rate, intersection over union, and the number of buildings extracted. With the improved CNN model and no requirement of further post-processing, we have generated building maps for the United States with an average processing time less than one minute for an area of size ∼ 56 km 2 . The quality of extracted buildings and processing time demonstrated the proposed CNN-based framework fits the need of building extraction at scale.

show abstract

Section: Pre-trained Modelsmentioning

confidence: 99%