Zhaorui Gu

Robust and automatic cell detection and segmentation from microscopic images of non-setae phytoplankton species

Saliency-based marker-controlled watershed method was proposed to detect and segment phytoplankton cells from microscopic images of non-setae species. This method first improved IG saliency detection method by combining saturation feature with colour and luminance feature to detect cells from microscopic images uniformly and then produced effective internal and external markers by removing various specific noises in microscopic images for efficient performance of watershed segmentation automatically. The authors built the first benchmark dataset for cell detection and segmentation, including 240 microscopic images across multiple phytoplankton species with pixel-wise cell regions labelled by a taxonomist, to evaluate their method. They compared their cell detection method with seven popular saliency detection methods and their cell segmentation method with six commonly used segmentation methods. The quantitative comparison validates that their method performs better on cell detection in terms of robustness and uniformity and cell segmentation in terms of accuracy and completeness. The qualitative results show that their improved saliency detection method can detect and highlight all cells, and the following marker selection scheme can remove the corner noise caused by illumination, the small noise caused by specks, and debris, as well as deal with blurred edges.

Automatic plankton image classification combining multiple view features via multiple kernel learning

BackgroundPlankton, including phytoplankton and zooplankton, are the main source of food for organisms in the ocean and form the base of marine food chain. As the fundamental components of marine ecosystems, plankton is very sensitive to environment changes, and the study of plankton abundance and distribution is crucial, in order to understand environment changes and protect marine ecosystems. This study was carried out to develop an extensive applicable plankton classification system with high accuracy for the increasing number of various imaging devices. Literature shows that most plankton image classification systems were limited to only one specific imaging device and a relatively narrow taxonomic scope. The real practical system for automatic plankton classification is even non-existent and this study is partly to fill this gap.ResultsInspired by the analysis of literature and development of technology, we focused on the requirements of practical application and proposed an automatic system for plankton image classification combining multiple view features via multiple kernel learning (MKL). For one thing, in order to describe the biomorphic characteristics of plankton more completely and comprehensively, we combined general features with robust features, especially by adding features like Inner-Distance Shape Context for morphological representation. For another, we divided all the features into different types from multiple views and feed them to multiple classifiers instead of only one by combining different kernel matrices computed from different types of features optimally via multiple kernel learning. Moreover, we also applied feature selection method to choose the optimal feature subsets from redundant features for satisfying different datasets from different imaging devices. We implemented our proposed classification system on three different datasets across more than 20 categories from phytoplankton to zooplankton. The experimental results validated that our system outperforms state-of-the-art plankton image classification systems in terms of accuracy and robustness.ConclusionsThis study demonstrated automatic plankton image classification system combining multiple view features using multiple kernel learning. The results indicated that multiple view features combined by NLMKL using three kernel functions (linear, polynomial and Gaussian kernel functions) can describe and use information of features better so that achieve a higher classification accuracy.

An imaging-inspired no-reference underwater color image quality assessment metric

Abstract Underwater color image quality assessment (IQA) plays an important role in analysis and applications of underwater imaging as well as image processing algorithms. This paper presents a new metric inspired by the imaging analysis on underwater absorption and scattering characteristics, dubbed the CCF. This metric is feature-weighted with a combination of colorfulness index, contrast index and fog density index, which can quantify the color loss caused by absorption, the blurring caused by forward scattering and the foggy caused by backward scattering, respectively. Then multiple linear regression is used to calculate three weighted coefficients. A new underwater image database is built to illustrate the performance of the proposed metric. Experimental results show a strong correlation between the proposed metric and mean opinion score (MOS). The proposed CCF metric outperforms many of the leading atmospheric IQA metrics, and it can effectively assess the performance of underwater image enhancement and image restoration methods.

Underwater Inherent Optical Properties Estimation Using a Depth Aided Deep Neural Network

Underwater inherent optical properties (IOPs) are the fundamental clues to many research fields such as marine optics, marine biology, and underwater vision. Currently, beam transmissometers and optical sensors are considered as the ideal IOPs measuring methods. But these methods are inflexible and expensive to be deployed. To overcome this problem, we aim to develop a novel measuring method using only a single underwater image with the help of deep artificial neural network. The power of artificial neural network has been proved in image processing and computer vision fields with deep learning technology. However, image-based IOPs estimation is a quite different and challenging task. Unlike the traditional applications such as image classification or localization, IOP estimation looks at the transparency of the water between the camera and the target objects to estimate multiple optical properties simultaneously. In this paper, we propose a novel Depth Aided (DA) deep neural network structure for IOPs estimation based on a single RGB image that is even noisy. The imaging depth information is considered as an aided input to help our model make better decision.

Underwater image sharpness assessment based on selective attenuation of color in the water

Due to the dramatic effect of scattering caused by underwater medium, blur commonly occurs in underwater images. Moreover, colors are frequently attenuated in underwater situations, in response to the effect that water selectively absorbs light of different wavelengths. Thus, this color loss becomes one of the key features of underwater images. Most present blur/sharpness evaluation algorithms, which are based on grayscale images, are able to perform well on normal overwater images. However, they could not assess underwater images sufficiently because the color distortion has been neglected. This paper proposes a method to measure the underwater image sharpness based on selective attenuation of color in the water. In addition, in order to evaluate the performance of our method, we exploit an underwater image database. Tests on LIVE blur database and underwater image database demonstrate that our method correlates better with subjective human quality evaluations compared with the other methods.

Unsupervised pixel-wise classification for Chaetoceros image segmentation

Abstract Chaetoceros is a dominant genus of marine planktonic diatoms with worldwide distribution. Due to the difficulty of extracting setae from Chaetoceros images, automatic segmentation of Chaetoceros is still a challenging task. In this paper, we address this difficult task by regarding the whole segmentation process as unsupervised pixel-wise classification without human participation. First, we automatically produce positive (object) and negative (background) samples for follow-up training, by combining the advantages of two image processing algorithms: Grayscale Surface Direction Angle Model (GSDAM) for extracting setae information and Canny for detecting cell edges from low-contrast and strong-noisy microscopic images. Second, we develop pixel-wise training by using the produced samples in the training process of Deep Convolutional Neural Network (DCNN). At last, the trained DCNN is used to label other pixels into object and background for final segmentation. We compare our method with eight mainstream segmentation approaches: Otsu’s thresholding, Canny, Watershed, Mean Shift, gPb-owt-ucm, Normalized Cut, Efficient Graph-based method and GSDAM. To objectively evaluate segmentation results, we apply six well-known evaluation indexes. Experimental results on a new Chaetoceros image dataset with human labelled ground truth show that our method outperforms the eight mainstream segmentation methods in terms of both quantitative and qualitative evaluation.

A practical system of fish size measurement

This paper proposes a practical system of the fish size measurement which is effective and simple. Because the fry we measured is very small, the main technical problem when we measure the fish size is to measure accurately without hurting them. Although the live fish can be measured directly or after anesthesia, the fry will be damaged. So we present the system of fish size measurement without any contact to them. The portable system consists of a USB camera fixed on a rod with a adjustable height and the camera will be connected to a computer for image capturing, two lights with adjustable intensity that will be used for illumination, and an image processing software we developed that will be used to calibrate and calculate the fish length and width. Finally, the actual length and width data will be recorded in an excel. Experiment and analysis are presented in this paper, which will verify the practicability and effectiveness of the system.

Simulation of stochastic resonance in underwater laser communication

Due to the light attenuation caused by absorption and scattering of sea water, the performance of underwater optical communication system is limited. Usually, the optical signals will be decayed severely in hundreds of meters and the waveforms are submerged in noise. To improve the signal-to-noise ratio (SNR), for traditional methods, most of the noise can be suppressed by filters based on classic linear theory. However, the signal will also be relatively attenuated by the filters. In this research, the nonlinear signal processing method, named as stochastic resonance (SR), is introduced to enhance the underwater laser communication signal. A SR system is simulated to evaluate the feasibility and effectiveness of weak signal detection in underwater laser communication. The influence of the noise to the bit-error-rate(BER) and the parameter optimization are simulated and discussed. It is proved that, the SR system has advantage over the low-pass filter(LPF) and has promising prospect in the signal enhancement in underwater laser communication.

A design of underwater wireless laser communication system based on PPM modulating method

Underwater laser communication has been developed in the field of deep-sea exploration, also become a highlight of modern communication technology research. It has the edges of hidden, safe, non-contact and rapid mobility. By now, most data exchange and transmission between underwater sensors and devices are achieved by RS-232 serial ports communication. But it is not able to meet the demands of present underwater communication due to the limitations of transmitting rate and distance. For the first version of our laser system the signals flowing through the RS-232 ports had not been modulated. Therefore, the bit error rate of our previous work was adversely high. In this paper, we progress our system by introducing in a novel signal modulating module, which mainly consists of FPGAs and PPM method. Considering with the effect of absorption and scattering, a series of experiments are performed in this paper: unmodulated and modulated underwater laser communication experiments under different visibility water. By comparing the bit error rate results, we suggest that the whole underwater laser communication system is compact, efficient and reliable. Finally, we based on multiple of visibility, the theoretical communication distance of the system can be estimated.

Automatic searching of fish from underwater images via shape matching

To conserve and manage fish stocks in fisheries, people monitor the growth and quantities of fish by collected underwater videos and images. However, the efficiency of finding fish from videos and images is limited by manual operation. In this paper, we propose an automatic method to improve the efficiency of finding fish by shape matching. In our method, a segmentation method, Segmentation by Aggregating Superpixels (SAS), is used to partition all objects from every underwater image. To search for fish from these partitioned objects, Inner-Distance Shape Context (IDSC) is applied as a shape matching method. The proposed method can overcome the disturbing of specific underwater environment, such as blurring and fishs deformation in images. Experimental result shows that the method is effective to search for fish from underwater images and it can be widely applied to fish surveillance in fisheries.