Underwater Image Enhancement Using Improved CNN Based Defogging

Due to refraction, absorption, and scattering of light by suspended particles in water, underwater images are characterized by low contrast, blurred details, and color distortion. In this paper, a fusion algorithm to restore and enhance underwater images is proposed. It consists of a color restoration module, an end-to-end defogging module and a brightness equalization module. In the color restoration module, a color balance algorithm based on CIE Lab color model is proposed to alleviate the effect of color deviation in underwater images. In the end-to-end defogging module, one end is the input image and the other end is the output image. A CNN network is proposed to connect these two ends and to improve the contrast of the underwater images. In the CNN network, a sub-network is used to reduce the depth of the network that needs to be designed to obtain the same features. Several depth separable convolutions are used to reduce the amount of calculation parameters required during network training. The basic attention module is introduced to highlight some important areas in the image. In order to improve the defogging network’s ability to extract overall information, a cross-layer connection and pooling pyramid module are added. In the brightness equalization module, a contrast limited adaptive histogram equalization method is used to coordinate the overall brightness. The proposed fusion algorithm for underwater image restoration and enhancement is verified by experiments and comparison with previous deep learning models and traditional methods. Comparison results show that the color correction and detail enhancement by the proposed method are superior.

Research on the Correlation between Breeding Environment and Activity of Yellow Feather Broilers Based on the Multichromatic Aberration Model

Broiler behavior is closely related to the breeding environment. Therefore, studying broiler behavior helps breeding farm workers to better carry out welfare breeding. In the breeding environment of yellow feather broilers, temperature, humidity, and ammonia concentration are the main factors that affect the behavior of the broilers. This study used a multichromatic aberration model to process the color images of yellow feather broilers to extract the activity feature of the broilers at different periods, utilized the Cb component of YCbCr color model and the b component of Lab color model to remove background litter in images, and employed the Q component of YIQ color model to remove the feeders and the drinkers from the image. The segmented images were constructed into an accumulator to generate a heat map of yellow feather broilers’ activity. Then, the correlation between the activity and the temperature and humidity index (THI) and the correlation between the activity and ammonia concentration were explored. The experiment found that the activity of the broilers was significantly positively correlated with ammonia concentration ( P < 0.05 ), indicating that the activity of yellow feather broilers increased with ammonia concentration ascending. Besides, the THI in the broiler chamber had a significant positive correlation with the ammonia data ( P < 0.01 ), indicating that when the THI in the broiler chamber increases, the ammonia concentration rises. The research provides a direction for exploring the impact of THI and ammonia concentration on the performance of yellow feather broilers. At the same time, it provides a theoretical basis for the early warning and judgment of broiler breeding by farm workers in the future.

Исследование изменений в спектре поглощения современных хлоринсодержащих препаратов для фотодинамической терапии и метиленового синего в результате воздействия светодиодного излучения с длиной волны 656± 10 nm

The changes in absorption spectra of aqueous solutions of modern chlorine-containing photosensitizing preparations "Revixan" (Areal, Russia) and "Chloderm" (Chloderm, Russia) depending on the intensity of LED radiation with wavelength of 656 ± 10 nm and exposure time were studied in spectral range 600-700 nm. The parameters of the CIE Lab color model of the image of "Revixan" aqueous solution before and after LED exposure were investigated. The changes in absorption spectra of aqueous solutions of methylene blue with different initial concentrations arising after exposure to LED radiation with intensity of 180 ± 20 mW/cm2 were studied in the spectral range 400-900 nm. It was shown that the impact of LED radiation changes the absorption spectra of the studied preparations and increases the parameter L (lightness) of the CIE Lab color model for "Revixan". An increase in the LED radiation intensity and exposure time leads to a decrease in absorption for "Revixan" and "Chloderm" in spectral range 600-700 nm and to a shift of the peaks of absorption bands lying in this range towards a longer wavelength. The impact of LED radiation on aqueous solutions of methylene blue leads to a decrease in their absorption in spectral range 400-900 nm.

Color models of interference images of thin stratified objects in optical microscopy

Algorithms for the analysis of polychromatic interference patterns in images of thin stratified objects in optical microscopy are considered. The algorithms allow one to measure the thin-film optical thickness. A measurement method based on the comparison of colors of the interference image under study and a numerically simulated image is discussed. We discuss a mathematical model for the calculation and numerical simulation of interference patterns and algorithms for interference pattern processing. Color comparison in an RGB color model is described and limitations of such a method are shown. The feasibility of using a Lab color model is shown and algorithms of interference color comparison in this model are presented. Results of application of the presented algorithms to measuring the optical thickness of red blood cells in a blood smear are discussed. The estimation of the error and robustness of the proposed algorithms is conducted.

Image Colorization

The article discusses one of the latest ways to colorize a black and white image using deep learning methods. For colorization, a convolutional neural network with a large number of layers (Deep convolutional) is used, the architecture of which includes a ResNet model. This model was pre-trained on images of the ImageNet dataset. A neural network receives a black and white image and returns a colorized color. Since, due to the characteristics of ResNet, an input multiple of 255 is received, a program was written that, using frames, enlarges the image for the required size. During the operation of the neural network, the CIE Lab color model is used, which allows to separate the black and white component of the image from the color. For training the neural network, the Place 365 dataset was used, containing 365 different classes, such as animals, landscape elements, people, and so on. The training was carried out on the Nvidia GTX 1080 video card. The result was a trained neural network capable of colorizing images of any size and format. As example we had a speed of 0.08 seconds and an image of 256 by 256 pixels in size. In connection with the concept of the dataset used for training, the resulting model is focused on the recognition of natural landscapes and urban areas.

Assessment the quality of apricots in the process of drying with neural networks and support vector machines

The paper presents an analysis of the assessment the quality of apricots during the drying process using two types of classifires: ANNs and SVMs. The quality of apricots is categorized in three classes according to the color and b-carotene content through the process of drying. The classification is made by using ‘CIE Lab’ color model and spectral characteristics in the VIS range. Neural networks are BPN and PNN, and classifiers are kernel and linear SVM. The spectral characteristics are pre-processed with SNV, MSC, First derivative and PCA. According to the results for color features, BPN and SVM with “rbf” kernel have the best performance while PNN has the worst performance. When using spectral characteristics the BPN network performs well: eavg = 4.1% and emax = 12.1% but the SVM linear (eavg = 3.4%, emax =5.3%) and SVM with “rbf” kernel (eavg = 2.4%, emax =5.2%) classifiers have better results. As a conclusion, it could be said that classifiers using spectral features perform well with errors at about 2-5%. Classification with color features is an alternative method, which is less complex, cheaper and with acceptable errors.

The Application of Computer Image Analysis Based on Textural Features for the Identification of Barley Kernels Infected with Fungi of the Genus Fusarium

The aim of this study was to develop discrimination models based on textural features for the identification of barley kernels infected with fungi of the genus Fusarium and healthy kernels. Infected barley kernels with altered shape and discoloration and healthy barley kernels were scanned. Textures were computed using MaZda software. The kernels were classified as infected and healthy with the use of the WEKA application. In the case of RGB, Lab and XYZ color models, the classification accuracies based on 10 selected textures with the highest discriminative power ranged from 95 to 100%. The lowest result (95%) was noted in XYZ color model and Multi Class Classifier for the textures selected using the Ranker method and the OneR attribute evaluator. Selected classifiers were characterized by 100% accuracy in the case of all color models and selection methods. The highest number of 100% results was obtained for the Lab color model with Naive Bayes, LDA, IBk, Multi Class Classifier and J48 classifiers in the Best First selection method with the CFS subset evaluator.

Using the pixel information value parameter in an entropy image analysis criterion

The work is devoted to the development and research of the entropy criterion of image analysis on its corresponding to the method of RLE-compression. The Lab color model and the CIEDE1976 color estimation metric have been analyzed. The pixel information importance parameter has been introduced, which is based on the above described metric and allows us to estimate the importance of the adjacent pixel taking into consideration the information it introduces in relation to the previous pixel. The modified entropy image analysis criterion has been developed taking into account the pixel’s information importance parameter. The adequacy of the proposed criterion has been checked on the sample of standard test images and the feasibility of its use has been proved.

Silhouette index for determining optimal k-means clustering on images in different color models

Clustering process is an essential part of the image processing. Its aim to group the data according to having the same attributes or similarities of the images. Consequently, determining the number of the optimum clusters or the best (well-clustered) for the image in different color models is very crucial. This is because the cluster validation is fundamental in the process of clustering and it reflects the split between clusters. In this study, the k-means algorithm was used on three colors model: CIE Lab, RGB and HSV and the clustering process made up to k clusters. Next, the Silhouette Index (SI) is used to the cluster validation process, and this value is range between 0 to 1 and the greater value of SI illustrates the best of cluster separation. The results from several experiments show that the best cluster separation occurs when k=2 and the value of average SI is inversely proportional to the number of k cluster for all color model. The result shows in HSV color model the average SI decreased 14.11% from k = 2 to k = 8, 11.1% in HSV color model and 16.7% in CIE Lab color model. Comparisons are also made for the three color models and generally the best cluster separation is found within HSV, followed by the RGB and CIE Lab color models.