Smart Imaging Analysis Techniques for Bioprocesses in Liquid Environment
Mayar Aly Mohammed Atteya;
Abstract
Nowadays, many developments allow the replacement of fish oil ingredients with the products yield from the biotechnological process of marine microalgae, which is used for the ω-3 production to large amounts. ω-3 fatty acids are used in protection against heart and cancer diseases.
For the improvement of biotechnological processes, an enhanced in-line imaging technology is strongly needed in order to track the morphological changes of the microalgae cells, which are important for the process yield, while the cells must be kept in a certain age.
As the microalgae cells live in submerged environment, the size parameter can differentiate between the biological cells, grain stones, and air bubbles. Indeed, particle size recognition is an important task in different applications those applied on microscopic imaging.
This thesis presents different methods for enhancing the automation of particles recognition, counting, and classifying using a recently developed SOPAT-System (Smart On-line Particle Analysis Technology), a photo-optical image acquisition device for in-situ microscopic imaging.
We have proposed three segmentation methods, namely, template matching, edge detection and multi-thresholding, and iterative contrast enhancement and image adaptation, to accurately identify the microalgae cells of the image. These methods are applied on different datasets as synthetic and real microscopic images of microalgae (low and high contrast).
The template matching method is applied on all datasets and approved that the segmentation results don’t have accurate values. The second method, edge detection and multi-thresholding based segmentation, includes image de-noising using Haar wavelet transform, image binarization using edge detection and multi-thresholding methods, image enhancement using morphological operations, and watershed transform, alternatively, circular Hough transform for touched cells separation. This method proves that the microalgae particles could be recognized correctly with accuracy reaches up to 99 %. But, it is limited to high contrast images.
The third one, iterative contrast enhancement and image adaptation based segmentation is applied on all datasets. It includes image de-noising, image normalization by Contrast Limited Histogram Equalization method, image enhancement by morphological operations, a region of interest extraction using an integral filter, and active contour method, or, fuzzy c-mean, alternatively, Otsu’s thresholding, circular Hough transform, or, watershed transform as segmentation step. The segmentation result of this method reaches up to 99 %.
Furthermore, for differentiating between the biological cells and others, the combination of cell’s size and cell’s texture analysis are measured for accurate classification based on segmentation process. The experimental results prove that the microalgae particles can be classified correctly with the accuracy reaches up to 100% compared to reference values.
For the improvement of biotechnological processes, an enhanced in-line imaging technology is strongly needed in order to track the morphological changes of the microalgae cells, which are important for the process yield, while the cells must be kept in a certain age.
As the microalgae cells live in submerged environment, the size parameter can differentiate between the biological cells, grain stones, and air bubbles. Indeed, particle size recognition is an important task in different applications those applied on microscopic imaging.
This thesis presents different methods for enhancing the automation of particles recognition, counting, and classifying using a recently developed SOPAT-System (Smart On-line Particle Analysis Technology), a photo-optical image acquisition device for in-situ microscopic imaging.
We have proposed three segmentation methods, namely, template matching, edge detection and multi-thresholding, and iterative contrast enhancement and image adaptation, to accurately identify the microalgae cells of the image. These methods are applied on different datasets as synthetic and real microscopic images of microalgae (low and high contrast).
The template matching method is applied on all datasets and approved that the segmentation results don’t have accurate values. The second method, edge detection and multi-thresholding based segmentation, includes image de-noising using Haar wavelet transform, image binarization using edge detection and multi-thresholding methods, image enhancement using morphological operations, and watershed transform, alternatively, circular Hough transform for touched cells separation. This method proves that the microalgae particles could be recognized correctly with accuracy reaches up to 99 %. But, it is limited to high contrast images.
The third one, iterative contrast enhancement and image adaptation based segmentation is applied on all datasets. It includes image de-noising, image normalization by Contrast Limited Histogram Equalization method, image enhancement by morphological operations, a region of interest extraction using an integral filter, and active contour method, or, fuzzy c-mean, alternatively, Otsu’s thresholding, circular Hough transform, or, watershed transform as segmentation step. The segmentation result of this method reaches up to 99 %.
Furthermore, for differentiating between the biological cells and others, the combination of cell’s size and cell’s texture analysis are measured for accurate classification based on segmentation process. The experimental results prove that the microalgae particles can be classified correctly with the accuracy reaches up to 100% compared to reference values.
Other data
| Title | Smart Imaging Analysis Techniques for Bioprocesses in Liquid Environment | Other Titles | طرق ذكية لتحليل صور التطور الحيوي في وسيط سائل | Authors | Mayar Aly Mohammed Atteya | Issue Date | 2016 |
Recommend this item
Similar Items from Core Recommender Database
Items in Ain Shams Scholar are protected by copyright, with all rights reserved, unless otherwise indicated.