Search

Multi-class segmentation of unlabelled living cells in time-lapse light microscopy images is challenging due to the temporal behaviour and changes in cell life cycles and the complexity of these images. The deep-learning-based methods achieved promising outcomes and remarkable success in single- and multi-class medical and microscopy image segmentation. The main objective of this study is to develop a hybrid deep-learning-based categorical segmentation and classification method for living HeLa cells in reflected light microscopy images. A symmetric simple U-Net and three asymmetric hybrid convolution neural networks—VGG19-U-Net, Inception-U-Net, and ResNet34-U-Net—were proposed and mutually compared to find the most suitable architecture for multi-class segmentation of our datasets. The inception module in the Inception-U-Net contained kernels with different sizes within the same layer to extract all feature descriptors. The series of residual blocks with the skip connections in each ResNet34-U-Net’s level alleviated the gradient vanishing problem and improved the generalisation ability. The m-IoU scores of multi-class segmentation for our datasets reached 0.7062, 0.7178, 0.7907, and 0.8067 for the simple U-Net, VGG19-U-Net, Inception-U-Net, and ResNet34-U-Net, respectively. For each class and the mean value across all classes, the most accurate multi-class semantic segmentation was achieved using the ResNet34-U-Net architecture (evaluated as the m-IoU and Dice metrics).

We developed a method of image preprocessing based on the information entropy, namely, on the information contribution made by each individual pixel to the whole image or to image’s part (i.e., a Point Information Gain; PIG). An idea of the PIG calculation is that an image background remains informatively poor, whereas objects carry relevant information. In one calculation, this method preserves details, highlights edges, and decreases random noise. This paper describes optimization and implementation of the PIG calculation on graphical processing units (GPU) to overcome a high computational burden.

Restocking programmes of different fish species have been implemented worldwide. However, the survival of hatchery-reared fish after release to riverine ecosystems is at a very low level. One of the reasons for the high mortality rate of post-released fish is their modified swimming behaviour due to the hatchery rearing practice. To investigate one of the possible causes for modified swimming behaviour, Acipenser baerii larvae were exposed to surface- and bottom-feeding applications with day and night light regimes in a factorial design. We also analysed the effect of 5 and 10 days of starvation after different feeding applications on sturgeon swimming behaviour. The surface-feeding application was previously expected to promote the frequent Siberian sturgeon swim up to the mid- and top-water layers in our rearing facilities. However, our results indicated that the modified behaviour of the Siberian sturgeon in our study was caused by fish starvation and a possible predator-free environment rather than by the method of feed application or the day/night light regimes. These results may be used to improve the implementation of restocking programmes either through modified hatchery rearing practice or the training of foraging skills with predator stimuli.

Fish feed delivery is one of the challenges which fish farmers encounter daily. The main aim of the feeding process is to ensure that every fish is provided with sufficient feed to maintain desired growth rates. The properties of fish feed pellet, such as water stability, degree of swelling or floating time, are critical traits impacting feed delivery. Some considerable effort is currently being made with regard to the replacement of fish meal and fish oil with other sustainable alternative raw materials (i.e., plant or insect-based) with different properties. The main aim of this study is to investigate the motion and residence time distribution (RTD) of two types of solid feed pellets with different properties in a cylindrical fish tank. After experimental identification of material and geometrical properties of both types of pellets, a detailed 3D computational fluid dynamics (CFD) study for each type of pellets is performed. The mean residence time of pellets injected at the surface of the fish tank can differ by up to 75% depending on the position of the injection. The smallest residence time is when the position is located at the center of the liquid surface (17 s); the largest is near the edge of the tank (75 s). The maximum difference between the two studied types of pellets is 25% and it increases with positions closer to the center of the tank. The maximum difference for positions along the perimeter at 3/4 tank radius is 8%; the largest residence times are observed at the opposite side of the water inlet. Based on this study, we argue that the suitability of different solid feed pellets for aquaculture systems with specific fish can be determined, and eventually the pellet composition (formula) as well as the injection position can be optimized.

The main aim of this study was to develop a new objective method for evaluating the impacts of different diets on the live fish skin using image-based features. In total, one-hundred and sixty rainbow trout (Oncorhynchus mykiss) were fed either a fish-meal based diet (80 fish) or a 100% plant-based diet (80 fish) and photographed using consumer-grade digital camera. Twenty-three colour features and four texture features were extracted. Four different classification methods were used to evaluate fish diets including Random forest (RF), Support vector machine (SVM), Logistic regression (LR) and k-Nearest neighbours (k-NN). The SVM with radial based kernel provided the best classifier with correct classification rate (CCR) of 82% and Kappa coefficient of 0.65. Although the both LR and RF methods were less accurate than SVM, they achieved good classification with CCR 75% and 70% respectively. The k-NN was the least accurate (40%) classification model. Overall, it can be concluded that consumer-grade digital cameras could be employed as the fast, accurate and non-invasive sensor for classifying rainbow trout based on their diets. Furthermore, these was a close association between image-based features and fish diet received during cultivation. These procedures can be used as non-invasive, accurate and precise approaches for monitoring fish status during the cultivation by evaluating diet’s effects on fish skin.

We generalize the point information gain (PIG) and derived quantities, i.e., point information gain entropy (PIE) and point information gain entropy density (PIED), for the case of the Rényi entropy and simulate the behavior of PIG for typical distributions. We also use these methods for the analysis of multidimensional datasets. We demonstrate the main properties of PIE/PIED spectra for the real data with the examples of several images and discuss further possible utilizations in other fields of data processing.