Previous Image Research Articles

Unsupervised water scene dehazing network using multiple scattering model.

In water scenes, where hazy images are subject to multiple scattering and where ideal data sets are difficult to collect, many dehazing methods are not as effective as they could be. Therefore, an unsupervised water scene dehazing network using atmospheric multiple scattering model is proposed. Unlike previous image dehazing methods, our method uses the unsupervised neural network and the atmospheric multiple scattering model and solves the problem of difficult acquisition of ideal datasets and the effect of multiple scattering on the image. In our method, in order to embed the atmospheric multiple scattering model into the unsupervised dehazing network, the unsupervised dehazing network uses four branches to estimate the scene radiation layer, transmission map layer, blur kernel layer and atmospheric light layer, the hazy image is then synthesized from the four output layers, minimizing the input hazy image and the output hazy image, where the output scene radiation layer is the final dehazing image. In addition, we constructed unsupervised loss functions which applicable to image dehazing by prior knowledge, i.e., color attenuation energy loss and dark channel loss. The method has a wide range of applications, with haze being thick and variable in marine, river and lake scenes, the method can be used to assist ship vision for target detection or forward road recognition in hazy conditions. Through extensive experiments on synthetic and real-world images, the proposed method is able to recover the details, structure and texture of the water image better than five advanced dehazing methods.

3-D Q-coda attenuation structure at Mt. Etna (Italy)

Summary Three dimensional attenuation images of Mt. Etna volcano obtained by the analysis of Q-coda from local volcano-tectonic earthquakes are presented in this work. Seismic sources are confined inside the Etna structure with a maximum focal depth of 35 km below the sea level. The space distribution of the attenuation values was calculated by using 3-D weighting functions derived by the sensitivity kernels of Pacheco and Snieders (2006) and approximated by a polynomial interpolation, represented in the maps by using a back- projection method. Data were analyzed in four bands with central frequency placed at 1.5, 3, 6 and 12 Hz respectively. We observed a frequency dependence of Q-coda with values that range from 55 at 1.5 Hz to 218 at 12 Hz. Q-coda space distribution in the Etna area shows almost uniformity in the average attenuation in the first 35 km below the surface. The images were derived with a resolution of 5 km. We observe as one of our main conclusions that Q-coda attenuation space anomalies are correlated with the areas of highest structural heterogeneities and are distributed along the well-known tectonic structures which characterize the crust in Mt. Etna region. Previous and numerous velocity and attenuation images describing the structure of Mt. Etna support our main conclusion: high Q-coda volumes almost coincide with the zones marked by high velocity and relative low total attenuation for direct waves.

Exploring nine simultaneously occurring transients on April 12th 1950

Nine point sources appeared within half an hour on a region within sim 10 arcmin of a red-sensitive photographic plate taken in April 1950 as part of the historic Palomar Sky Survey. All nine sources are absent on both previous and later photographic images, and absent in modern surveys with CCD detectors which go several magnitudes deeper. We present deep CCD images with the 10.4-m Gran Telescopio Canarias, reaching brightness r sim 26 mag, that reveal possible optical counterparts, although these counterparts could equally well be just chance projections. The incidence of transients in the investigated photographic plate is far higher than expected from known detection rates of optical counterparts to e.g. flaring dwarf stars, Fast Radio Bursts, Gamma Ray Bursts or microlensing events. One possible explanation is that the plates have been subjected to an unknown type of contamination producing mainly point sources with of varying intensities along with some mechanism of concentration within a radius of sim 10 arcmin on the plate. If contamination as an explanation can be fully excluded, another possibility is fast (t <0.5 s) solar reflections from objects near geosynchronous orbits. An alternative route to confirm the latter scenario is by looking for images from the First Palomar Sky Survey where multiple transients follow a line.

Road Characteristics Detection Based on Joint Convolutional Neural Networks with Adaptive Squares

The importance of road characteristics has been highlighted, as road characteristics are fundamental structures established to support many transportation-relevant services. However, there is still huge room for improvement in terms of types and performance of road characteristics detection. With the advantage of geographically tiled maps with high update rates, remarkable accessibility, and increasing availability, this paper proposes a novel simple deep-learning-based approach, namely joint convolutional neural networks (CNNs) adopting adaptive squares with combination rules to detect road characteristics from roadmap tiles. The proposed joint CNNs are responsible for the foreground and background image classification and various types of road characteristics classification from previous foreground images, raising detection accuracy. The adaptive squares with combination rules help efficiently focus road characteristics, augmenting the ability to detect them and provide optimal detection results. Five types of road characteristics—crossroads, T-junctions, Y-junctions, corners, and curves—are exploited, and experimental results demonstrate successful outcomes with outstanding performance in reality. The information of exploited road characteristics with location and type is, thus, converted from human-readable to machine-readable, the results will benefit many applications like feature point reminders, road condition reports, or alert detection for users, drivers, and even autonomous vehicles. We believe this approach will also enable a new path for object detection and geospatial information extraction from valuable map tiles.

Developed but close to nature: the image of Finland in the National Geographic Magazine from the 1900s to the 2010s

ABSTRACT The present article focuses on the image of Finland in the National Geographic Magazine between 1905 and 2013. The study contributes to the research on national images by answering the following questions: a) how are Finland and the Finns represented in the photojournalist articles in the magazine, b) how has the image of Finland changed over the decades, and c) what kind of cultural, social, and political meanings are conveyed through the image(s)? The research material consists of 37 English written articles including in total 250 photographs and other images. The results show four overlapping but still distinctive phases in the thematic transformation of the image: Finland as a part of the diverse Russian Empire, Finland as a progressive but traditional European nation, Finland as the opposite to the Soviet Union, and Finland as a country of wild nature. The findings are discussed in light of the previous national image research.

Hue‐shifting for accurate and precise quantification of biochemical substances using diagnostic test strips

AbstractSome test strips have high feasibility for the quick and approximate quantification of the biochemical substances in samples. Recently, test strips became more precise and accurate by involving the quantitative observation of coloration and regression modelling. Their accuracy and precision can be significantly improved by hue‐shifting the colours of the original sample and those of the standards. The standard colour images included in the instructions of test strips for the detection of four urinary and two salivary biochemical substances were used. Images of glucose‐loaded test strips for urinary glucose detection were also prepared. Using the red‐green‐blue (RGB) colour image, the author prepared 11 hue‐shifted colour images by changing the hue of the original RGB colour image at intervals of 30°. From each of the 12 colour images, 10 greyscale images indicating values of the intensity of RGB, cyan, magenta, yellow, key black, L*, a* and b* were prepared. In regression modelling, the additional 110 greyscale images improved the values of the coefficient of determination and normalised mean absolute error in the quantification of protein, specific gravity, urobilinogen and glucose in urine, as well as of alcohol and nitric oxide in saliva. Further analyses showed that the hue‐shifting technique yielded new greyscale images that were more significantly related to the values of the biochemical variables in question than any previous greyscale images. This improvement was realised by the most significant linear measures for the description of concentrations of biochemical substances, and synergetic effects between significant greyscale images.

Diabetic Oculomotor Nerve Palsy Displaying Enhancement of the Oculomotor Nerve in the Orbit and Cavernous Sinus on MRI

Introduction: Imaging data were scarce on diabetic oculomotor nerve palsy (ONP). Our study explored the MRI features and their clinical implications for diabetic ONP. Methods: Fifty-nine patients with a clinical diagnosis of diabetic ONP were recruited from our department between January 2015 and December 2019. Orbital MRI was retrospectively analyzed, and follow-up scans were obtained for 5 patients. Based on the ocular motor nerve palsy scale, the difference in the scores on the first and last hospital days was defined as the improvement score and was used to assess the treatment effects in all. Results: Thirty-eight (64.41%) patients presented thickening and enhancement of the cavernous segment and inferior division of the intraorbital segment of the ipsilateral oculomotor nerve, with the cisternal segment spared in all. After complete resolution of symptoms, follow-up MRI in 5 patients revealed that the enhancement was less obvious compared with the previous images. 6 patients in the enhancement group and 4 patients in the nonenhancement group were treated with 80 mg of methylprednisolone. Significant differences were not detected in the median improvement scores between patients with and those without corticosteroid use (p = 0.240). Conclusion: Thickening and enhancement of the unilateral oculomotor nerve were common imaging findings in diabetic ONP, and they persisted after complete resolution of symptoms in some patients. The cavernous segment and the inferior division of the intraorbital segment were simultaneously involved, and the cisternal segment was often spared. Refraining from corticosteroids was recommended even with nerve enhancement.

Image Region Prediction from Thermal Videos Based on Image Prediction Generative Adversarial Network

Various studies have been conducted on object detection, tracking, and action recognition based on thermal images. However, errors occur during object detection, tracking, and action recognition when a moving object leaves the field of view (FOV) of a camera and part of the object becomes invisible. However, no studies have examined this issue so far. Therefore, this article proposes a method for widening the FOV of the current image by predicting images outside the FOV of the camera using the current image and previous sequential images. In the proposed method, the original one-channel thermal image is converted into a three-channel thermal image to perform image prediction using an image prediction generative adversarial network. When image prediction and object detection experiments were conducted using the marathon sub-dataset of the Boston University-thermal infrared video (BU-TIV) benchmark open dataset, we confirmed that the proposed method showed the higher accuracies of image prediction (structural similarity index measure (SSIM) of 0.9839) and object detection (F1 score (F1) of 0.882, accuracy (ACC) of 0.983, and intersection over union (IoU) of 0.791) than the state-of-the-art methods.

Photoacoustic Tomography Image Restoration With Measured Spatially Variant Point Spread Functions.

The spatial resolution of photoacoustic tomography (PAT) can be characterized by the point spread function (PSF) of the imaging system. Due to the tomographic detection geometry, the PAT image degradation model could be generally described by using spatially variant PSFs. Deconvolution of the PAT image with these PSFs could restore image resolution and recover object details. Previous PAT image restoration algorithms assume that the degraded images can be restored by either a single uniform PSF, or some blind estimation of the spatially variant PSFs. In this work, we propose a PAT image restoration method to improve image quality and resolution based on experimentally measured spatially variant PSFs. Using photoacoustic absorbing microspheres, we design a rigorous PSF measurement procedure, and successfully acquire a dense set of spatially variant PSFs for a commercial cross-sectional PAT system. A pixel-wise PSF map is further obtained by employing a multi-Gaussian-based fitting and interpolation algorithm. To perform image restoration, an optimization-based iterative restoration model with two kinds of regularizations is proposed. We perform phantom and in vivo mice imaging experiments to verify the proposed method, and the results show significant image quality and resolution improvement.

Single Image Super Resolution Using Multiple Re-Evaluation Process

In this paper, we proposed improved single image super resolution using multiple re-evaluation Process model for use in various image processing fields. The proposed method generates the first super resolution using the input image, and analyzes the change for each region by comparing the features of previous image and super resolution result. According to the analyzed features, the feature map for generate n-th super resolution was selected for improved detail. After then, next generate super resolution using previous super resolution result as input image. This process is repeated for final result. The existing single image super resolution method has two areas to be improved. First, it minimizes artifacts or staircases, which are unnecessary details that can be created during the super resolution process. Second, it is necessary to consider the input image because it affects the result depending on the quality of input image used in the super resolution process. Therefore, in order to minimize unnecessary details, the proposed method analyzed the feature map from the generated super resolution result and applied it according to the amount of change. In addition, aimed to gradually improve the input image to be used in the super resolution process by using the super resolution generated in the previous step. By comparing and evaluating the proposed method with the conventional single image super resolution method with PSNR and SSIM, it is improved by about 3%.

Impurities Detection in Intensity Inhomogeneous Edible Bird’s Nest (EBN) Using a U-Net Deep Learning Model

As an important export, cleanliness control on edible bird’s nest (EBN) is paramount. Automatic impurities detection is in urgent need to replace manual practices. However, effective impurities detection algorithm is yet to be developed due to the unresolved inhomogeneous optical properties of EBN. The objective of this work is to develop a novel U-net based algorithm for accurate impurities detection. The algorithm leveraged the convolution mechanisms of U-net for precise and localized features extraction. Output probability tensors were then generated from the deconvolution layers for impurities detection and positioning. The U-net based algorithm outperformed previous image processing-based methods with a higher impurities detection rate of 96.69% and a lower misclassification rate of 10.08%. The applicability of the algorithm was further confirmed with a reasonably high dice coefficient of more than 0.8. In conclusion, the developed U-net based algorithm successfully mitigated intensity inhomogeneity in EBN and improved the impurities detection rate.

Combined Gray Level Transformation Technique for Low Light Color Image Enhancement

Insufficient and poor lightning conditions affect the quality of videos and images captured by the camcorders. The low quality images decrease the performances of computer vision systems in smart traffic, video surveillance, and other imaging systems applications. In this paper, combined gray level transformation technique is proposed to enhance the less quality of illuminated images. This technique is composed of log transformation, power law transformation and adaptive histogram equalization process to improve the low light illumination image estimated using HIS color model. Finally, the enhanced illumination image is blended with original reflectance image to get enhanced color image. This paper shows that the proposed algorithm on various weakly illuminated images is enhanced better and has taken reduced computation time than previous image processing techniques.

Interferometric phase image denoising method via residual learning

The denoising of interferometric phase images attracts many researchers. Natural image denoising algorithms based on neural networks are often proposed in the development of deep learning methods. A neural network, In-CNN, is derived from an advanced natural image denoising network and proposed for interferometric phase image denoising. Preactivation and residual learning methods are combined and applied to the function of the neural network nodes. Considering the particularity of the interferometric phase image, we propose a neural network based on the rational application of the preactivation mode and feedforward mapping, which is different from previous natural image denoising networks. We also construct a training set for an interferometric phase image denoising neural network. We experimentally verify that our model performs better than state-of-the-art interferometric phase image denoising methods based on sparse representation and advanced natural image denoising networks. We discuss the complexity of traditional interferometric phase image denoising algorithms to demonstrate the efficiency of the proposed method.

Optimizing early cancer diagnosis and detection using a temporal subtraction technique

To optimize the early diagnosis and detection of lung cancer, computer-aided diagnostic (CAD) systems have been a useful tool for analyzing medical images. The temporal subtraction technique, which is a CAD system, performs the subtraction operation between the current image and the previous image on the same patient, and supports observation by emphasizing the temporal changes. However, the temporal subtraction technique for 3D images, such as thoracic CT images, has not yet been established. There is a need to develop efficient and highly accurate 3D nonrigid registration techniques to reduce subtraction artifacts. This study aims to develop a 3D nonrigid registration technique to establish a 3D temporal subtraction technique. In particular, we focus on the Finite Element Method, which is versatile, applicable to a wide range of fields, and capable of handling any shape. Our new method was examined on 46 clinical cases with multidetector row computed tomography images. As a result, the proposed method improved by 6.93% (p = 3.0 × 10−6) compared to the conventional methods in terms of the rate of reduction of artifacts, and the effectiveness was verified. Therefore, this study contributes to the literature on early detection and treatment.

Neural Twins Talk and Alternative Calculations

Inspired by how the human brain employs more neural pathways when increasing the focus on a subject, we introduce a novel twin cascaded attention model that outperforms a state-of-the-art image captioning model that was originally implemented using one channel of attention for the visual grounding task. Visual grounding ensures the existence of words in the caption sentence that are grounded into a particular region in the input image. After a deep learning model is trained on visual grounding task, the model employs the learned patterns regarding the visual grounding and the order of objects in the caption sentences, when generating captions. We report the results of our experiments in three image captioning tasks on the COCO dataset. The results are reported using standard image captioning metrics to show the improvements achieved by our model over the previous image captioning model. The results gathered from our experiments suggest that employing more parallel attention pathways in a deep neural network leads to higher performance. Our implementation of Neural Twins Talk (NTT) is publicly available at: https://github.com/zanyarz/NeuralTwinsTalk .

Use of Time-Series NDWI to Monitor Emerging Archaeological Sites: Case Studies from Iraqi Artificial Reservoirs

Over the last 50 years, countries across North Africa and the Middle East have seen a significant increase in dam construction which, notwithstanding their benefits, have endangered archaeological heritage. Archaeological surveys and salvage excavations have been carried out in threatened areas in the past, but the formation of reservoirs often resulted in the permanent loss of archaeological data. However, in 2018, a sharp fall in the water level of the Mosul Dam reservoir led to the emersion of the archaeological site of Kemune and allowed for its brief and targeted investigation. Reservoir water level change is not unique to the Mosul Dam, but it is a phenomenon affecting most of the artificial lakes of present-day Iraq. However, to know in advance which sites will be exposed due to a decrease in water level can be a challenging task, especially without any previous knowledge, field investigation, or high-resolution satellite image. Nonetheless, by using time-series medium-resolution satellite images, combined to obtain spectral indexes for different years, it is possible to monitor “patterns” of emerging archaeological sites from three major Iraqi reservoirs: Mosul, Haditha and Hamrin lake. The Normalised Difference Water Index (NDWI), generated from annual composites of Landsat and Sentinel-2 images, allow us to distinguish between water bodies and other land surfaces. When coupled with a pixel analysis of each image, the index can provide a mean for highlighting whether an archaeological site is submerged or not. Moreover, using a zonal histogram algorithm in QGIS over polygon shapefiles that represent a site surface, it is possible to assess the area of a site that has been exposed over time. The same analyses were carried out on monthly composites for the year 2018, to assess the impact of monthly variation of the water level on the archaeological sites. The results from both analyses have been visually evaluated using medium-resolution true colour images for specific years and locations and with 3 m resolution Planetscope images for 2018. Understanding emersion “patterns” of known archaeological sites provides a useful tool for targeted rescue excavation, while also expanding the knowledge of the post-flooding impact on cultural heritage in the regions under study.

OtoPair: Combining Right and Left Eardrum Otoscopy Images to Improve the Accuracy of Automated Image Analysis

The accurate diagnosis of otitis media (OM) and other middle ear and eardrum abnormalities is difficult, even for experienced otologists. In our earlier studies, we developed computer-aided diagnosis systems to improve the diagnostic accuracy. In this study, we investigate a novel approach, called OtoPair, which uses paired eardrum images together rather than using a single eardrum image to classify them as ‘normal’ or ‘abnormal’. This also mimics the way that otologists evaluate ears, because they diagnose eardrum abnormalities by examining both ears. Our approach creates a new feature vector, which is formed with extracted features from a pair of high-resolution otoscope images or images that are captured by digital video-otoscopes. The feature vector has two parts. The first part consists of lookup table-based values created by using deep learning techniques reported in our previous OtoMatch content-based image retrieval system. The second part consists of handcrafted features that are created by recording registration errors between paired eardrums, color-based features, such as histogram of a* and b* component of the L*a*b* color space, and statistical measurements of these color channels. The extracted features are concatenated to form a single feature vector, which is then classified by a tree bagger classifier. A total of 150-pair (300-single) of eardrum images, which are either the same category (normal-normal and abnormal-abnormal) or different category (normal-abnormal and abnormal-normal) pairs, are used to perform several experiments. The proposed approach increases the accuracy from 78.7% (±0.1%) to 85.8% (±0.2%) on a three-fold cross-validation method. These are promising results with a limited number of eardrum pairs to demonstrate the feasibility of using a pair of eardrum images instead of single eardrum images to improve the diagnostic accuracy.

A Brief Analysis on the Construction of Subject Image in the Self-Documentary "Small Talk"

Taiwanese producer Huang Hui-chen pointed the camera at herself and her mother, trying to achieve the purpose of communication and reconciliation with her mother through the means of documentary. The difference from previous self-recorded images is that the conflicts of the subject construction in “Small Talk” is no longer limited to the aspects of “people with people”, “person with family”, “parents with children”, but it constructed a subject image of an ordinary family that is a collection of various social issues such as homosexuality, domestic violence, sexual assault, and aboriginal Taiwanese culture etc.

Designing the Virus

‘Designing the Virus’ brings together my ongoing research across the histories of science and design in two specific areas: practices of visualizing viruses for both scientific and public communication; and design in response to risk. In 2020, these areas intersected in a way that was impossible, as a researcher, to ignore, when a medical illustration of coronavirus released by the United States Centers for Disease Control and Prevention (CDC) took on life in a way that no previous scientific image has done, as a frequently cut-and-pasted, remixed, and broadcast signal not only of the danger posed by the virus, but of the pandemic itself. By summer 2020, the ‘spiky blob’ coronavirus illustration had grown prevalent in the visual culture of numerous countries across the globe. This essay, written in August 2020 in part as a ‘timestamp’ of the period, questions what it means for a biomedical image to become an icon for a global crisis. I analyse the image from interdisciplinary angles of the history of graphic design for public health communication and scientific image-making practices, draw upon published interviews with the CDC’s medical illustrators, and build on recent research in disaster studies that critiques the notion of ‘natural disaster’. The medical illustration of the coronavirus presents an unusual case in the history of public risk communication, as a stand-alone scientific image that has come to act as a piece of wordless risk communication. I argue that this image, and its widespread dissemination in the US (and beyond), ‘implicitly reinforces a specific position within the politics of risk that have been unfolding in the US during the pandemic: that the overriding threat is the virus itself, divorced from the social, political, and environmental factors that shape how lives across the globe are affected by this pathogen’.

Does the access transport mode affect visitors' satisfaction in a World Heritage City? The case of Valparaiso, Chile

The inscription of a city as a World Heritage Site (WHS) acts as a catalyst for the tourist promotion. The brand creates a previous positive image of the city generating some tourists' expectations that might be difficult to fulfil. The aim of this paper is to shed some light on how different transport modes -private car, rental cars, train, bus and other transport modes- could mediate on the satisfaction construct. The methodology is based on a fuzzy hybrid multi-criteria method. Our results show that rental cars visitors experience the minimum satisfaction in comparison with those who visit the city by bus who are the most satisfied. The findings can be strategically used by Valparaiso authorities and destination marketing managers to promote Valparaíso as a green destination WHS.