311 research outputs found
Atlanta’s Desegregation Era Social Studies Curriculum: An Examination of Georgia History Textbooks
No full textAuthor accepted manuscript version of a chapter published in:
Bohan, C. H. & Randolph, P. (2012). Desegregation era social studies curriculum: An examination of Georgia History textbooks. Chapter seven in C. Woyshner and C. H. Bohan (Eds.) Histories of social studies and race, 1890–2000. (pp. 135−158). New York: Palgrave MacMillan.</p
Network properties data and code used in "Ecological plasticity governs ecosystem services in multilayer networks".
No full textCode and network properties data used in the analyses presented in "Ecological plasticity governs ecosystem services in multilayer networks". Further information can be requested of the author David A. Bohan ([email protected])
Selective Compression of Medical Images via Intelligent Segmentation and 3D-SPIHT Coding
No full textABSTRACT SELECTIVE COMPRESSION OF MEDICAL IMAGES VIA INTELLIGENT SEGMENTATION AND 3D-SPIHT CODING by Bohan Fan The University of Wisconsin-Milwaukee, 2018 Under the Supervision of Professor Zeyun Yu With increasingly high resolutions of 3D volumetric medical images being widely used in clinical patient treatments, efficient image compression techniques have become in great demand due to the cost in storage and time for transmission. While various algorithms are available, the conflicts between high compression rate and the downgraded quality of the images can partially be harmonized by using the region of interest (ROI) coding technique. Instead of compressing the entire image, we can segment the image by critical diagnosis zone (the ROI zone) and background zone, and apply lossless compression or low compression rate to the former and high compression rate to the latter, without losing much clinically important information. In this thesis, we explore a medical image transmitting process that utilizes a deep learning network, called 3D-Unet to segment the region of interest area of volumetric images and 3D-SPIHT algorithm to encode the images for compression, which can be potentially used in medical data sharing scenario. In our experiments, we train a 3D-Unet on a dataset of spine images with their label ground truth, and use the trained model to extract the vertebral bodies of testing data. The segmented vertebral regions are dilated to generate the region of interest, which are subject to the 3D-SPIHT algorithm with low compress ratio while the rest of the image (background) is coded with high compress ratio to achieve an excellent balance of image quality in region of interest and high compression ratio elsewhere
Selective Compression of Medical Images via Intelligent Segmentation and 3D-SPIHT Coding
No full textABSTRACT
SELECTIVE COMPRESSION OF MEDICAL IMAGES VIA INTELLIGENT SEGMENTATION AND 3D-SPIHT CODING
by
Bohan Fan
The University of Wisconsin-Milwaukee, 2018
Under the Supervision of Professor Zeyun Yu
With increasingly high resolutions of 3D volumetric medical images being widely used in clinical patient treatments, efficient image compression techniques have become in great demand due to the cost in storage and time for transmission. While various algorithms are available, the conflicts between high compression rate and the downgraded quality of the images can partially be harmonized by using the region of interest (ROI) coding technique. Instead of compressing the entire image, we can segment the image by critical diagnosis zone (the ROI zone) and background zone, and apply lossless compression or low compression rate to the former and high compression rate to the latter, without losing much clinically important information.
In this thesis, we explore a medical image transmitting process that utilizes a deep learning network, called 3D-Unet to segment the region of interest area of volumetric images and 3D-SPIHT algorithm to encode the images for compression, which can be potentially used in medical data sharing scenario. In our experiments, we train a 3D-Unet on a dataset of spine images with their label ground truth, and use the trained model to extract the vertebral bodies of testing data. The segmented vertebral regions are dilated to generate the region of interest, which are subject to the 3D-SPIHT algorithm with low compress ratio while the rest of the image (background) is coded with high compress ratio to achieve an excellent balance of image quality in region of interest and high compression ratio elsewhere
Doctor Imitator: Hand-Radiography-based Bone Age Assessment by Imitating Scoring Methods
No full textBone age assessment is challenging in clinical practice due to the
complicated bone age assessment process. Current automatic bone age assessment
methods were designed with rare consideration of the diagnostic logistics and
thus may yield certain uninterpretable hidden states and outputs. Consequently,
doctors can find it hard to cooperate with such models harmoniously because it
is difficult to check the correctness of the model predictions. In this work,
we propose a new graph-based deep learning framework for bone age assessment
with hand radiographs, called Doctor Imitator (DI). The architecture of DI is
designed to learn the diagnostic logistics of doctors using the scoring methods
(e.g., the Tanner-Whitehouse method) for bone age assessment. Specifically, the
convolutions of DI capture the local features of the anatomical regions of
interest (ROIs) on hand radiographs and predict the ROI scores by our proposed
Anatomy-based Group Convolution, summing up for bone age prediction. Besides,
we develop a novel Dual Graph-based Attention module to compute
patient-specific attention for ROI features and context attention for ROI
scores. As far as we know, DI is the first automatic bone age assessment
framework following the scoring methods without fully supervised hand
radiographs. Experiments on hand radiographs with only bone age supervision
verify that DI can achieve excellent performance with sparse parameters and
provide more interpretability.Comment: Original Title: "Doctor Imitator: A Graph-based Bone Age Assessment
Framework Using Hand Radiographs" @inproceedings{chen2020doctor,
title={Doctor imitator: A graph-based bone age assessment framework using
hand radiographs}, author={Chen, Jintai and Yu, Bohan and Lei, Biwen and
Feng, Ruiwei and Chen, Danny Z and Wu, Jian}, booktitle={MICCAI}, year={2020}
Change of the spatial and temporal pattern of ecological vulnerability: A case study on Cheng-Yu urban agglomeration, Southwest China
No full textChina's urban economy has developed rapidly over the decades, and the Cheng-Yu urban agglomeration has become one of China's four typical urban agglomerations, with a large population and a high level of economic development. However, the conflict between humans and the environment is becoming increasingly prominent together with economic development. In order to protect the ecological environment in urban areas, thus scientific understanding and assessment of ecological vulnerability are beneficial to establishing regional conservation measures, and serve as a key means to maintain environmental health. Based on the “Sensitivity-Resilience-Pressure” (SRP) model, this study considered remote sensing, geographic and statistical data to construct an evaluation system for regional ecological vulnerability. In addition, the coupled AHP (Analytic hierarchy process)-Entropy weighting model was proposed to obtain the weight of each evaluation indicator and analyze the spatio-temporal distribution characteristics of the ecological vulnerability of the study area during 2000–2020. The changes and the divergence pattern were depicted by the transfer matrix, dynamic degree and spatial auto-correlation. The results indicated that the ecological vulnerability of Cheng-Yu urban agglomeration is mainly mild and moderate, with an overall high distribution in Chongqing and Chengdu, while low in the central and north zone (e.g., Ziyang, Mianyang). It is consistent with the distribution of HH (High-High) and L-L (Low-Low) clusters, respectively, having a significant positive spatial correlation. In particular, the severely vulnerable area increased from 7059 km2 in 2000 to 23553 km2 in 2020, with an increased rate of 233.66 %. Combining the transfer matrix and dynamic degree, it was found that the ecological environment underwent a rapid deterioration followed by a slow recovery. This study provides a scientific reference for the ecological policy making which serves sustainable urban development
Analysis on the Negative Impact of AI Development on Employment and Its Countermeasures
No full textWhile benefiting society, the technological progress of artificial intelligence (AI) has also brought a rising number of unemployed people and breeded polarization in income distribution by threatening the low-skill and labor-intensive industry. To solve the negative impact of AI, policies about the taxation and subsidy on AI and the income-supporting program can be proposed. However, neither of them will work well to achieve sustainable social development. In the long run, technological progress will not be influenced by government policies, and capital will find its own path to a rapid growth. Income-supporting programs are short-term solutions, being ineffective and not sustainable. Based on the literature collected, the author came up with two practical methods to deal with the negative impact brought by AI to employment: the industrial relocation as a short-term solution and the reframing of the education system as a long-term solution
Optics And Computer Vision For Biomedical Applications
No full textBioengineering is at the cross sections of biology, clinical technology, electrical engineering, computer science and many other domains. The smooth translation of domain technologies to clinics is not just about accuracy and practicality of the technology. It also has to take into account the accessibility (cost and portability), the patients’ comfort and the ease to adapt into the workflow of medical professionals. The dissertation will explore three projects, (1) portable and low-cost near infrared florescence imaging system on mobile phone platform, (2) computer aided diagnosis software for diagnosing chronical kidney disease based on optical coherence tomography (OCT) images and (3) the tracking and localization of hand-held medical imaging probe. These projects aim to translate and adapt modern computation hardware, data analysis models and computer vision technologies to solve and refine clinical diagnosis applications. The dissertation will discuss how the translation, tradeoffs and refinement of those technologies can bring a positive impact on the accuracy, ease of conduct, accessibility and patients’ comfort to the clinical applications
Estimating interactions of functional brain connectivity by Hidden Markov models
No full textThe brain activity reflected by functional magnetic resonance imaging (fMRI) is temporally organized as a combination of sensory inputs from environment and its own spontaneous activity. However, temporal patterns of brain activity in a large number of subjects remain unclear. We propose a regularized hidden Markov model (HMM) to estimate dynamic functional connectivity among distributed brain regions and discover repeating connectivity patterns from resting-state functional connectivity across a group of subjects. We found that functional brain connectivity are hierarchically organized and exhibit three repeated patterns across subjects with attention deficit hyperactivity disorder (ADHD). We have examined the temporal characteristics of functional connectivity by its occupancy. And we validated our method by comparing the classification performance with state-of-the-art methods using the same dataset. Experimental results show that our method can improve the classification performance compared to other functional connectivity modelling methods
- …
