×
  • Select the area you would like to search.
  • ACTIVE INVESTIGATIONS Search for current projects using the investigator's name, institution, or keywords.
  • EXPERTS KNOWLEDGE BASE Enter keywords to search a list of questions and answers received and processed by the ADNI team.
  • ADNI PDFS Search any ADNI publication pdf by author, keyword, or PMID. Use an asterisk only to view all pdfs.
Principal Investigator  
Principal Investigator's Name: Michael Coopman
Institution: University of Texas at Arlington
Department: Mathematics
Country:
Proposed Analysis: The ADNI data will be used to test a method for medical image registration. In particular, we plan to replicate and refine the method performed in a joint paper between Tsinghua University and University of Texas at Arlington. This method uses Voxelmorph, a framework for performing image registration through deep learning. Typically, the loss function for registration is based on a similarity term and a smoothness term. The method outlined replaces the smoothness term with the Laplacian of the transformation. Then, by defining a control function to the Laplacian, a more consistent transformation can be created. According to the paper, this resulted in a better mean Dice score than Voxelmorph registration methods. The paper referred to is the following: Yongpei Zhu, Zicong Zhou Sr., Guojun Liao Sr., and Kehong Yuan "New loss functions for medical image registration based on VoxelMorph", Proc. SPIE 11313, Medical Imaging 2020: Image Processing, 113132E (10 March 2020); https://doi.org/10.1117/12.2550030
Additional Investigators  
Investigator's Name: Guojun Liao
Proposed Analysis: The images being requested will be used to test a method for medical image registration. In particular, we plan to replicate and refine the method performed in a joint paper between Tsinghua University and University of Texas at Arlington. Typically, the loss function of registration is based on a similarity term and a smoothness term. The method outlined replaces the smoothness term with the Laplacian of the transformation. Then, by defining a control function to the Laplacian, a more consistent transformation can be created. According to the paper, this resulted in a better mean Dice score than other Voxelmorph registration methods. The paper referred to is the following: Yongpei Zhu, Zicong Zhou Sr., Guojun Liao Sr., and Kehong Yuan "New loss functions for medical image registration based on VoxelMorph", Proc. SPIE 11313, Medical Imaging 2020: Image Processing, 113132E (10 March 2020); https://doi.org/10.1117/12.2550030