-
GlacierNet2: A Hybrid Multi-Model Learning Architecture for Alpine Glacier Map**
Authors:
Zhiyuan Xie,
Umesh K. Haritashya,
Vijayan K. Asari,
Michael P. Bishop,
Jeffrey S. Kargel,
Theus H. Aspiras
Abstract:
In recent decades, climate change has significantly affected glacier dynamics, resulting in mass loss and an increased risk of glacier-related hazards including supraglacial and proglacial lake development, as well as catastrophic outburst flooding. Rapidly changing conditions dictate the need for continuous and detailed observations and analysis of climate-glacier dynamics. Thematic and quantitat…
▽ More
In recent decades, climate change has significantly affected glacier dynamics, resulting in mass loss and an increased risk of glacier-related hazards including supraglacial and proglacial lake development, as well as catastrophic outburst flooding. Rapidly changing conditions dictate the need for continuous and detailed observations and analysis of climate-glacier dynamics. Thematic and quantitative information regarding glacier geometry is fundamental for understanding climate forcing and the sensitivity of glaciers to climate change, however, accurately map** debris-cover glaciers (DCGs) is notoriously difficult based upon the use of spectral information and conventional machine-learning techniques. The objective of this research is to improve upon an earlier proposed deep-learning-based approach, GlacierNet, which was developed to exploit a convolutional neural-network segmentation model to accurately outline regional DCG ablation zones. Specifically, we developed an enhanced GlacierNet2 architecture thatincorporates multiple models, automatic post-processing, and basin-level hydrological flow techniques to improve the map** of DCGs such that it includes both the ablation and accumulation zones. Experimental evaluations demonstrate that GlacierNet2 improves the estimation of the ablation zone and allows a high level of intersection over union (IOU: 0.8839) score. The proposed architecture provides complete glacier (both accumulation and ablation zone) outlines at regional scales, with an overall IOU score of 0.8619. This is a crucial first step in automating complete glacier map** that can be used for accurate glacier modeling or mass-balance analysis.
△ Less
Submitted 29 July, 2022; v1 submitted 6 April, 2022;
originally announced April 2022.
-
Skin Cancer Segmentation and Classification with NABLA-N and Inception Recurrent Residual Convolutional Networks
Authors:
Md Zahangir Alom,
Theus Aspiras,
Tarek M. Taha,
Vijayan K. Asari
Abstract:
In the last few years, Deep Learning (DL) has been showing superior performance in different modalities of biomedical image analysis. Several DL architectures have been proposed for classification, segmentation, and detection tasks in medical imaging and computational pathology. In this paper, we propose a new DL architecture, the NABLA-N network, with better feature fusion techniques in decoding…
▽ More
In the last few years, Deep Learning (DL) has been showing superior performance in different modalities of biomedical image analysis. Several DL architectures have been proposed for classification, segmentation, and detection tasks in medical imaging and computational pathology. In this paper, we propose a new DL architecture, the NABLA-N network, with better feature fusion techniques in decoding units for dermoscopic image segmentation tasks. The NABLA-N network has several advances for segmentation tasks. First, this model ensures better feature representation for semantic segmentation with a combination of low to high-level feature maps. Second, this network shows better quantitative and qualitative results with the same or fewer network parameters compared to other methods. In addition, the Inception Recurrent Residual Convolutional Neural Network (IRRCNN) model is used for skin cancer classification. The proposed NABLA-N network and IRRCNN models are evaluated for skin cancer segmentation and classification on the benchmark datasets from the International Skin Imaging Collaboration 2018 (ISIC-2018). The experimental results show superior performance on segmentation tasks compared to the Recurrent Residual U-Net (R2U-Net). The classification model shows around 87% testing accuracy for dermoscopic skin cancer classification on ISIC2018 dataset.
△ Less
Submitted 24 April, 2019;
originally announced April 2019.
-
Advanced Deep Convolutional Neural Network Approaches for Digital Pathology Image Analysis: a comprehensive evaluation with different use cases
Authors:
Md Zahangir Alom,
Theus Aspiras,
Tarek M. Taha,
Vijayan K. Asari,
TJ Bowen,
Dave Billiter,
Simon Arkell
Abstract:
Deep Learning (DL) approaches have been providing state-of-the-art performance in different modalities in the field of medical imagining including Digital Pathology Image Analysis (DPIA). Out of many different DL approaches, Deep Convolutional Neural Network (DCNN) technique provides superior performance for classification, segmentation, and detection tasks. Most of the task in DPIA problems are s…
▽ More
Deep Learning (DL) approaches have been providing state-of-the-art performance in different modalities in the field of medical imagining including Digital Pathology Image Analysis (DPIA). Out of many different DL approaches, Deep Convolutional Neural Network (DCNN) technique provides superior performance for classification, segmentation, and detection tasks. Most of the task in DPIA problems are somehow possible to solve with classification, segmentation, and detection approaches. In addition, sometimes pre and post-processing methods are applied for solving some specific type of problems. Recently, different DCNN models including Inception residual recurrent CNN (IRRCNN), Densely Connected Recurrent Convolution Network (DCRCN), Recurrent Residual U-Net (R2U-Net), and R2U-Net based regression model (UD-Net) have proposed and provide state-of-the-art performance for different computer vision and medical image analysis tasks. However, these advanced DCNN models have not been explored for solving different problems related to DPIA. In this study, we have applied these DCNN techniques for solving different DPIA problems and evaluated on different publicly available benchmark datasets for seven different tasks in digital pathology including lymphoma classification, Invasive Ductal Carcinoma (IDC) detection, nuclei segmentation, epithelium segmentation, tubule segmentation, lymphocyte detection, and mitosis detection. The experimental results are evaluated with different performance metrics such as sensitivity, specificity, accuracy, F1-score, Receiver Operating Characteristics (ROC) curve, dice coefficient (DC), and Means Squired Errors (MSE). The results demonstrate superior performance for classification, segmentation, and detection tasks compared to existing machine learning and DCNN based approaches.
△ Less
Submitted 19 April, 2019;
originally announced April 2019.
