WEBVTT
Kind: captions
Language: en

00:00:14.420 --> 00:00:16.860
Hello, ladies and gentlemen.

00:00:16.860 --> 00:00:18.680
My name is Syu-Jyun Peng.

00:00:18.680 --> 00:00:22.340
I come from Taipei Medical University

00:00:22.340 --> 00:00:25.400
It's my honor to introduce the topic.

00:00:25.400 --> 00:00:28.300
The title of the topic is

00:00:28.300 --> 00:00:34.900
Artificial Intelligence Image Analytics of Lesional and Non-lesional

00:00:34.900 --> 00:00:37.460
Neurological Disorders

00:00:37.460 --> 00:00:40.660
with magnetic resonance imaging (MRI)

00:00:40.660 --> 00:00:50.480
including its chemical stroke, cerebral AVM and Non-lesional epilepsy.

00:00:50.480 --> 00:00:54.100
This is my outline, including introduction,

00:00:54.100 --> 00:00:56.080
method and results,

00:00:56.080 --> 00:00:58.640
and conclusions.

00:00:58.640 --> 00:01:08.140
Magnetic Resonance Image (MRI) contains pathology-related information.

00:01:08.140 --> 00:01:17.280
Detection of MRI-based biomarkers is of diagnostic and therapeutic value.

00:01:17.280 --> 00:01:23.140
Accurate detection of this type of markers is challenging

00:01:23.140 --> 00:01:29.280
because they may not be directly discernible

00:01:29.280 --> 00:01:35.420
and some are even non-lesional on conventional MRI.

00:01:35.760 --> 00:01:43.420
Any characteristic of cerebral tissue can be objectively evaluated

00:01:43.420 --> 00:01:50.320
and represented a parameter of its biological,

00:01:50.320 --> 00:01:53.980
functional or structural organization.

00:01:53.980 --> 00:02:01.220
An imaging biomarker is a measurable parameter obtained with standard

00:02:01.220 --> 00:02:12.540
and advanced techniques to explore, quantify and represent a tissue specific property.

00:02:12.540 --> 00:02:22.660
These properties are extracted after applying to the acquired images different computational models

00:02:22.700 --> 00:02:26.560
and specific statistical processing.

00:02:26.560 --> 00:02:37.480
The parametric maps illustrate the spatial distribution and signal intensity in the analyzed tissue.

00:02:37.980 --> 00:02:44.880
Medical imaging is an important technique in diagnosis,

00:02:44.880 --> 00:02:53.440
treatment monitoring and prognosis of the therapeutic response of diseases.

00:02:53.440 --> 00:03:05.120
It is also a fundamental means for guiding many minimally invasive therapeutic procedures.

00:03:05.600 --> 00:03:14.020
Traditional radiological diagnosis is based on the integration

00:03:14.020 --> 00:03:23.240
and qualitative assessment of imaging findings obtained from conventional radiography,

00:03:23.240 --> 00:03:27.400
ultrasound, computed tomography and magnetic resonance imaging (MRI).

00:03:27.400 --> 00:03:35.920
Technology and engineering have improved to acquire a variety of information

00:03:35.920 --> 00:03:38.460
from medical imaging.

00:03:38.460 --> 00:03:46.620
The knowledge of pathological and physiological information of neurological diseases

00:03:46.620 --> 00:03:56.680
has also elevated the application of these new parameters, known as biomarkers.

00:03:56.680 --> 00:04:04.820
This dissertation presents different methods for finding lesional biomarkers of acute ischemic stroke

00:04:04.820 --> 00:04:10.780
and non-lesional ones of neocortical seizures.

00:04:10.780 --> 00:04:15.360
For lesional biomarkers of acute ischemic stroke,

00:04:15.360 --> 00:04:22.040
we proposed a computer-assisted segmentation and quantification method

00:04:22.040 --> 00:04:28.880
to depict cerebral infarct, white matter hyperintensities (WMH), and AVM.

00:04:29.060 --> 00:04:33.260
For non-lesional biomarkers of neocortical epilepsy,

00:04:33.260 --> 00:04:40.800
a fiber-labeled MRI template was transformed to each subject’s neuroanatomy

00:04:40.800 --> 00:04:51.220
to generate personalized atlases for objective and automated regional-of-interest demarcation.