Generative and discriminative model-based approaches to microscopic image restoration and segmentation

Abstract Image processing is one of the most important applications of recent machine learning (ML) technologies. Convolutional neural networks (CNNs), a popular deep learning-based ML architecture, have been developed for image processing applications. However, the application of ML to microscopic images is limited as microscopic images are often 3D/4D, that is, the image sizes can be very large, and the images may suffer from serious noise generated due to optics. In this review, three types of feature reconstruction applications to microscopic images are discussed, which fully utilize the recent advancements in ML technologies. First, multi-frame super-resolution is introduced, based on the formulation of statistical generative model-based techniques such as Bayesian inference. Second, data-driven image restoration is introduced, based on supervised discriminative model-based ML technique. In this application, CNNs are demonstrated to exhibit preferable restoration performance. Third, image segmentation based on data-driven CNNs is introduced. Image segmentation has become immensely popular in object segmentation based on electron microscopy (EM); therefore, we focus on EM image processing.

Download Full-text

An Efficient Image Segmentation Using Partial Differential Equation for Image Processing Applications

Journal of Medical Imaging and Health Informatics ◽

10.1166/jmihi.2021.3840 ◽

2021 ◽

Vol 11 (10) ◽

pp. 2538-2545

Author(s):

P. Geetha ◽

S. Nagarani

Keyword(s):

Differential Equation ◽

Mathematical Method ◽

Image Processing ◽

Partial Differential Equation ◽

Image Segmentation ◽

Image Restoration ◽

Geometric Analysis ◽

Stochastic Modelling ◽

Partial Differential ◽

Structural Work

Different processing of the images, such as the image captured, saved and retrieved from another use of the specific image, must be restructured in various ways in the process. More methods such as image restoration, picture segmentation, improvement of the picture etc can be used when processing images. Reconstructed in 3D picture 2D pictures are need to be proper. Including geometric wavelets and geometric analysis the structural work focused upon a variational and a selectable differential equation to test PDE’s which is a convergence of stochastic modelling and analysis of harmonics. This paper focuses primarily on the critical reviews of the image segmentation collection with the PDE application as a mathematical method and introduces the key tool of mathematics and techniques along with the literature-based analysis.

Download Full-text

Microscopic Image Segmentation of Chinese Herbal Medicine Based on Region Growing Algorithm

Advanced Materials Research ◽

10.4028/www.scientific.net/amr.756-759.4110 ◽

2013 ◽

Vol 756-759 ◽

pp. 4110-4115 ◽

Cited By ~ 4

Author(s):

Qing Liu ◽

Li Jun Zhang ◽

Xi Ping Liu

Keyword(s):

Image Segmentation ◽

Herbal Medicine ◽

Chinese Herbal Medicine ◽

Cell Structure ◽

Region Growing ◽

Microscopic Image ◽

Seed Selection ◽

Seed Point ◽

Microscopic Images ◽

Chinese Herbal

In order to effectively separate the target region of the microscopic image of Chinese Herbal Medicine (CHM), and lay the foundation for the subsequent image recognition processing, a microscopic image segmentation method of CHM by using region growing (RG) algorithm is put forward based on the characteristics of the plant microscopic images. Firstly, the CHM microscopic images with different cell structure are regarded as a multi-dimensional matrix to process and established seed label matrix. Secondly, in a given region threshold conditions, the different seed growth points are selected to segmented the different images. Finally, given a fixed growth points, the microscopic images are processed by choosing a different threshold. The experimental results show that CHM image segmentation threshold and seed selection decide the image target extraction. For different CHM images, according to a certain method, the better image segmentation results can be achieved in the case to obtain a suitable threshold value using image information and the seed point adjustment.

Download Full-text

An improved method of microscopic image segmentation

MATEC Web of Conferences ◽

10.1051/matecconf/201817601041 ◽

2018 ◽

Vol 176 ◽

pp. 01041

Author(s):

Zhang Feng Shou ◽

Dong Fang ◽

Liu Jian Ting ◽

Meng Xin

Keyword(s):

Genetic Algorithm ◽

Image Processing ◽

Image Segmentation ◽

Maximum Entropy ◽

Microscopic Image ◽

Two Dimensional ◽

Segmentation Method ◽

Improved Method ◽

Image Optimization ◽

Segmentation Image

In order to improve the effectiveness and accuracy of image processing in modern medical inspection, a segmentation image optimization algorithm of improved two-dimensional maximum entropy threshold based on genetic algorithm combined with mathematical morphology is proposed, in view of the microscopic cell images characteristic and the shortcomings of the traditional segmentation algorithm. Through theoretical analysis and contrast test, the segmentation method proposed is superior to the traditional threshold segmentation method in microscopic cell images, and the average segmentation time of the improved algorithm is 73% and 44% higher than the traditional two-dimensional maximum entropy threshold and the improved two-dimensional maximum entropy threshold.

Download Full-text

Convolutional Neural Network Combined with Half-Quadratic Splitting Method for Image Restoration

Journal of Sensors ◽

10.1155/2020/8813413 ◽

2020 ◽

Vol 2020 ◽

pp. 1-12

Author(s):

Ran Li ◽

Lin Luo ◽

Yu Zhang

Keyword(s):

Neural Network ◽

Convolutional Neural Network ◽

Image Restoration ◽

Computing Time ◽

Splitting Method ◽

Super Resolution ◽

Optimization Method ◽

Discriminative Learning ◽

Advantages And Disadvantages ◽

Model Based

Generally, there are mainly two methods to solve the image restoration task in low-level computer vision, i.e., the model-based optimization method and the discriminative learning method. However, these two methods have clear advantages and disadvantages. For example, it is flexible for the model-based optimization method to handle different problems, but large quantity of computing time is required for better performance. The discriminative learning approach has high computing efficiency, but the application scope is seriously limited by the fixed training model. It would be better to combine the advantages of these two methods. Luckily, with the variable splitting techniques, we insert the trained convolutional neural network (CNN) for denoising as one model to the model-based optimization method to solve other image restoration problems (e.g., deblurring and super-resolution). Final experimental results show that our denoising network is able to provide strong prior information for image restoration tasks. The image restoration effects can reach or approximate the most advanced algorithm in such three tasks as denoising, deblurring, and super-resolution. Moreover, the algorithm proposed in this paper is also the most competitive in terms of computational efficiency.

Download Full-text

Super-resolution Technologies for High-Quality Image Processing

The Journal of the Institute of Electrical Engineers of Japan ◽

10.1541/ieejjournal.130.297 ◽

2010 ◽

Vol 130 (5) ◽

pp. 297-300

Author(s):

Nobuyuki MATSUMOTO

Keyword(s):

Image Processing ◽

Super Resolution ◽

High Quality ◽

High Quality Image ◽

Quality Image

Download Full-text

Application of PDE model based on K-SVD in millimeter wave image restoration

Journal of Computer Applications ◽

10.3724/sp.j.1087.2012.00756 ◽

2013 ◽

Vol 32 (3) ◽

pp. 756-758

Author(s):

Li SHANG ◽

Pin-gang SU

Keyword(s):

Image Restoration ◽

Millimeter Wave ◽

Pde Model ◽

Model Based ◽

Wave Image

Download Full-text

The New Trend of State Estimation: From Model-Driven to Hybrid-Driven Methods

Sensors ◽

10.3390/s21062085 ◽

2021 ◽

Vol 21 (6) ◽

pp. 2085

Author(s):

Xue-Bo Jin ◽

Ruben Jonhson Robert RobertJeremiah ◽

Ting-Li Su ◽

Yu-Ting Bai ◽

Jian-Lei Kong

Keyword(s):

Kalman Filter ◽

State Estimation ◽

Future Development ◽

Data Driven ◽

Estimation Methods ◽

Multiple Model ◽

Development Trends ◽

Research Results ◽

Model Driven ◽

Model Based

State estimation is widely used in various automated systems, including IoT systems, unmanned systems, robots, etc. In traditional state estimation, measurement data are instantaneous and processed in real time. With modern systems’ development, sensors can obtain more and more signals and store them. Therefore, how to use these measurement big data to improve the performance of state estimation has become a hot research issue in this field. This paper reviews the development of state estimation and future development trends. First, we review the model-based state estimation methods, including the Kalman filter, such as the extended Kalman filter (EKF), unscented Kalman filter (UKF), cubature Kalman filter (CKF), etc. Particle filters and Gaussian mixture filters that can handle mixed Gaussian noise are discussed, too. These methods have high requirements for models, while it is not easy to obtain accurate system models in practice. The emergence of robust filters, the interacting multiple model (IMM), and adaptive filters are also mentioned here. Secondly, the current research status of data-driven state estimation methods is introduced based on network learning. Finally, the main research results for hybrid filters obtained in recent years are summarized and discussed, which combine model-based methods and data-driven methods. This paper is based on state estimation research results and provides a more detailed overview of model-driven, data-driven, and hybrid-driven approaches. The main algorithm of each method is provided so that beginners can have a clearer understanding. Additionally, it discusses the future development trends for researchers in state estimation.

Download Full-text