Joseph J. Bartlett

Joseph J. Bartlett

Auto-generated profile, based on publicly available information

Joseph J. Bartlett

Department of Biomedical Engineering and Graeme Clark Institute The University of Melbourne Victoria Australia

Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
29 July 2024 in IEEE Transactions on Pattern Analysis and Machine Intelligence
Verified
Journal Article
Closed access
TZTianyang Zhang
Tianyang Zhang
SZShaoming Zheng
Shaoming Zheng
JCJun Cheng
Jun Cheng
XJXi Jia
Xi Jia
JBJoseph Bartlett
Joseph Bartlett
JDJinming Duan
Jinming Duan
Data distribution gaps often pose significant challenges to the use of deep segmentation models. However, retraining models for each distribution is expensive and time-consuming. In clinical contexts, device-embedded algorithms and networks, typically unretrainable and unaccessable post-manufacture, exacerbate this issue. Generative translation methods offer a solution to mitigate the gap by transferring data across domains. However, existing methods mainly focus on intensity distributions while ignoring the gaps due to structure disparities. In this paper, we formulate a new image-to-image translation task to reduce structural gaps. We propose a simple, yet powerful Structure-Unbiased Adversarial (SUA) network which accounts for both intensity and structural differences between the training and test sets for segmentation. It consists of a spatial transformation block followed by an intensity distribution rendering module. The spatial transformation block is proposed to reduce the structural gaps between the two images. The intensity distribution rendering module then renders the deformed structure to an image with the target intensity distribution. Experimental results show that the proposed SUA method has the capability to transfer both intensity distribution and structural content between multiple pairs of datasets and is superior to prior arts in closing the gaps for improving segmentation.

9 June 2024 in Magnetic Resonance in Medicine
Verified
Journal Article
Open access
JBJoseph J. Bartlett
Joseph J. Bartlett
CDCatherine E. Davey
Catherine E. Davey
LJLeigh A. Johnston
Leigh A. Johnston
JDJinming Duan
Jinming Duan
Purpose: The aim of this study was to develop a model‐based deep learning architecture to accurately reconstruct fiber orientation distributions (FODs) from a reduced number of diffusion‐weighted images (DWIs), facilitating accurate analysis with reduced acquisition times.Methods: Our proposed architecture, Spherical Deconvolution Network (SDNet), performed FOD reconstruction by mapping 30 DWIs to fully sampled FODs, which have been fit to 288 DWIs. SDNet included DWI‐consistency blocks within the network architecture, and a fixel‐classification penalty within the loss function. SDNet was trained on a subset of the Human Connectome Project, and its performance compared with FOD‐Net, and multishell multitissue constrained spherical deconvolution.Results: SDNet achieved the strongest results with respect to angular correlation coefficient and sum of squared errors. When the impact of the fixel‐classification penalty was increased, we observed an improvement in performance metrics reliant on segmenting the FODs into the correct number of fixels.Conclusion: Inclusion of DWI‐consistency blocks improved reconstruction performance, and the fixel‐classification penalty term offered increased control over the angular separation of fixels in the reconstructed FODs.

4 June 2025 in Magnetic Resonance in Medicine
Verified
Journal Article
Open access
ITIgor Tyshchenko
Igor Tyshchenko
SLSimon Lévy
Simon Lévy
JBJoseph J. Bartlett
Joseph J. Bartlett
BTBahman Tahayori
Bahman Tahayori
YBYasmin Blunck
Yasmin Blunck
LJLeigh A. Johnston
Leigh A. Johnston
Purpose: The aim of the study was to conduct in vivo evaluation of population‐specific pulses in the context of the MP2RAGE sequence for brain imaging at 7T.Methods: Five clusters were identified in a cohort of 39 volunteers. One cluster centered around the average head shape and position, whereas four others were located toward the extremes of the population distribution. Additionally, the one‐size‐fits‐all solution was considered, using a standard universal pulse (UP) approach. Head shapes were characterized using lateral head breadth (HB), anterior–posterior head length (HL), and Y‐shift metrics. For each group, a 5kT‐points universal inversion pulse was computed and evaluated on four new test anatomies. A Python pipeline was integrated into the image reconstruction routine, using localizer scans to classify head shapes and positions. The pipeline selected one of five precomputed population‐specific pulses or defaulted to the generic UP without extending scan time.Results: The pipeline accurately classified head shapes and selected suitable pulses, enhancing the contrast of MP2RAGE images. Population‐specific pulses helped mitigate some of the performance loss associated with using a one‐size‐fits‐all UP, bringing performance closer to that of fully tailored solutions. This approach was particularly beneficial for individuals with smaller head sizes. However, the performance was worse in larger, more challenging head shapes.Conclusion: The novel head clustering and pulse selection pipeline facilitates the implementation of population‐specific pulses in clinical practice by allowing pulse selection tailored to each head shape and position without increasing scan time.