Diyasa, I Gede Susrama Mas (2019) Penentuan Abnormalitas Lintasan Pergerakan Spermatozoa Pada Video Mikroskopis Menggunakan Modifikasi Frame Difference Dan Regresi Linear. Doctoral thesis, Institut Technology Sepuluh Nopember.
Preview |
Text
07111160010002-Disertation.pdf Download (5MB) | Preview |
Abstract
Penelitian ini mengusulkan beberapa metode dalam deteksi, menghitung
jumlah dan pelacakan lintasan pergerakan spermatozoa berbasis video secara
cerdas. Tiga bagian besar yang diteliti, yaitu: penentuan abnormalitas morfologi
spermatozoa, abnormalitas motility spermatozoa yang terdiri dari modifikasi
background subtraction untuk penjejakan spermatozoa dan penentuan
abnormalitas pergerakan spermatozoa berdasarkan lintasan.
Pada bagian penentuan abnormalitas morfologi digunakan metode SVM
(Support Vector Machine) yang dibandingkan dengan metode K-NN (K-Nearest
Neighbour) untuk identifikasi abnormalitas pada bentuk kepala spermtozoa. Pada
bagian pelacakan kepala spermatozoa digunakan metode M-Frame Difference.
Pada bagian ekstraksi fitur untuk penentuan abnormalitas bentuk kepala
spermatozoa antara lain area, eccentricity dan ECD sesudah dilakukan BLOB
Analysis. Pada bagian kedua dengan memodifikasi beberapa algoritma
background subtraction untuk memisahkan objek sperma dari cairan semen.
Penelitian ini melalukan deteksi dan perhitungan spermatozoa yang bergerak pada
data video. Untuk melakukan deteksi pada sperma yang bergerak, metode Mixture
of Gaussian V2 background subtraction digunakan. Metode ini sesuai dalam
kasus deteksi sperma karena data sperma yang digunakan cenderung uni-modal.
Penelitian ini juga membandingkan metode background subtraction lainnya
dalam melakukan deteksi sperma.
Bagian ketiga dilakukan penentuan abnormalitas pergerakan berbasis
algoritma regresi linaer pada spermatozoa dalam semen, dari lintasan yang
terbentuk dianalisa normal tidaknya pergerakan sperma dalam semen. Dari hasil
percobaan yang dilakukan video data spermatozoa manusia, ternyata metode di
atas didapat posisi pergerakan spermatozoa hasil penjejakan dikenali bentuk
lintasannya berdasarkan rata-rata jarak posisinya terhadap garis regresi linier,
dengan threshold RMS sebesar 10 terdapat 10 spermatozoa progresif dan 4
spermatozoa non progresif
===============================================================================================
sections were examined, namely: determination of morphological
abnormalities of spermatozoa, abnormalities of spermatozoa motility which
consisted of the modification of background subtraction for tracking spermatozoa,
and determination of abnormalities of spermatozoa movement based on the
trajectory.
In the determination of morphological abnormalities, SVM (Support
Vector Machine) method is used which is then compared to the K-NN (K-Nearest
Neighbor) method to identify abnormalities in the spermatozoa's head shape. In
the tracking section of the spermatozoa head, the M-Frame Difference method
was used. Some extraction features performed to determine spermatozoa head
shape abnormalities include area, eccentricity, and ECD after BLOB Analysis.
The second part modified some background subtraction algorithms to separate
sperm objects from semen. This study detected and calculated moving
spermatozoa in video data. To detect the moving sperm, the Mixture of Gaussian
V2 background subtraction method is used. This method is suitable in the case of
sperm detection because sperm data tends to be uni-modal. This study also
compared other background subtraction methods for sperm detection.
The third part determined the movement abnormalities based on the linear
regression algorithm on spermatozoa in cement; the trajectory formed is analyzed
whether the movement of sperm in cement is normal or not. From the results of
experiments conducted on human spermatozoa video data, it turns out that the
above method obtained the position of spermatozoa tracking results identified by
the shape of the track based on the average distance of the linear regression line.
With an RMS threshold of 10, there are 10 progressive spermatozoa and 4 nonprogressive
spermatozoa.
| Item Type: | Thesis (Doctoral) |
|---|---|
| Additional Information: | RDE 681.413 Diy p-1 2019 |
| Uncontrolled Keywords: | M-Frame Difference Algorithm, Mixture of Gaussian V2, Regresi Linear, Lintasan Spermatozoa. |
| Subjects: | P Language and Literature > PN Literature (General) > PN1992.94 Video recordings--Production and direction. Q Science > QA Mathematics > QA278.2 Regression Analysis. Logistic regression |
| Divisions: | Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Electrical Engineering > 20001-(S3) PhD Thesis |
| Depositing User: | Diyasa I Gede Susrama Mas |
| Date Deposited: | 28 Mar 2022 02:45 |
| Last Modified: | 28 Mar 2022 02:45 |
| URI: | http://repository.its.ac.id/id/eprint/62216 |
Actions (login required)
 |
View Item |