Artificial Intelligence

Computational techniques on images with different applications

When:
Friday, August 24, 2018, 11:00am - 12:00pm PDTiCal
Where:
6th Floor Conf Rm (#689)
Type:
AI Seminar
Speaker:
Hengameh Mirzaalian
Video:
https://bluejeans.com/s/O9a8u
Description:
This talk is mainly an overview of my projects applied on medical images (e.g. retinal image, dermoscopic image,  CT, MRI, diffusion MRI, PET) and mars orbital images. The general goal was applying image-processing and machine-learning techniques to address different tasks like: data harmonization; image segmentation; image registration; quality assessment of data; high resolution image reconstruction; artifact detection; artifact correction; lesion tracking; landmark detection; landmark classification,  motion detection;  and object classification using supervised and semi supervised techniques. 
 
 
Bio:
I received my PhD  from Simon Fraser University, BC, Canada. My dissertation research focused on computational techniques for skin lesion tracking and classification. In Feb2012, I had an internship at Siemens Corporate Research and Technologies, on which I was applying shape analysis methods for vertebra segmentation on CT images. During my postdoc at Psychiatry Neuroimaging Laboratory and Computational Radiology Laboratory  (Dec2014- Feb2018), I was working on diffusion MRI data harmonization  and high resolution reconstruction of them.  I joined JPL on Feb2018 to design a pipeline to perform landmark classification task on mars orbital images. During my journey on working on different projects, I tried to learn more about  image-processing, feature-engineering, machine-learning, and deep-learning techniques to address different tasks on images with different modalities. 
 
« Return to Events