Visual Degradation Research Articles

SCREENING AND PARTIAL 16 S RRNA GENE SEQUENCING OF KERATINASE PRODUCING STENOTROPHOMONASMALTOPHILIA KARUNA5 ISOLATED FROM POULTRY WASTE

Keratinase is a protease which degrades the insoluble protein keratin that is largely found in feather waste. Fifty-seven soil samples were collected from a feather and poultry waste dumping sites. All these soil samples were subjected for enrichment in Whole feather meal medium containing 1% chicken feather as a sole source of Carbon, nitrogen and energy with pH 8.0 and incubated at 30°C on shaker condition (150 rpm) for 7 days. Out of 57 flasks only 21 showed visual feather degradation. Enriched sample from each flask was isolated on skimmed milk agar plates. Only 10 isolates demonstrated proteolytic activity by showing zone of clearance around the colony on skimmed milk agar medium. The above 10 bacterial isolates (Kar:01-12) were assayed for keratinase enzyme by growing them in Whole feather meal medium pH 8. Maximum keratinase production was demonstrated by isolate number Kar05 (41U/ml) at 30°C on shaker condition (150rpm) after 4 days of incubation. Promising isolate was identified as Stenotrophomonasmaltophila KARUNA5 by morphological, cultural, biochemical and 16S rRNA sequence analysis and was submitted to NCBI (Accession no: LC271188).

Visual and Physical Degradation of the Black and White Mosaic of a Roman Domus under Palazzo Valentini in Rome: A Preliminary Study.

Palazzo Valentini, the institutional head office of Città Metropolitana di Roma Capitale, stands in in a crucial position in the Roman archaeological and urban contexts, exactly between the Fora valley, Quirinal Hill slopes, and Campus Martius. It stands on a second-century A.D. complex to which belong, between other archeological remains, two richly decorated aristocratic domus. One of these buildings, the domus A, presents an outward porticoed room with a fourth-century AD central impluvium (open air part of the atrium designed to carry away rainwater) with a black/white tiled mosaic pavement, the preservation status of which is compromised by an incoherent degradation product that has caused gradual detachment of the mosaic tiles. To identify the product and determine the causes of degradation, samples of the product were taken and subjected to SEM-EDS, XRF, NMR, FT-IR and GC-MS analyses. The findings reported in this study can help restorers, archaeologists and conservation scientists in order to improve knowledge about the Roman mosaic, its construction phases, conservation problems and proper solutions.

No-Reference Quality Assessment for 3D Colored Point Cloud and Mesh Models

To improve the viewer's Quality of Experience (QoE) and optimize computer graphics applications, 3D model quality assessment (3D-QA) has become an important task in the multimedia area. Point cloud and mesh are the two most widely used digital representation formats of 3D models, the visual quality of which is quite sensitive to lossy operations like simplification and compression. Therefore, many related studies such as point cloud quality assessment (PCQA) and mesh quality assessment (MQA) have been carried out to measure the visual quality degradations of 3D models. However, a large part of previous studies utilize full-reference (FR) metrics, which indicates they can not predict the quality level with the absence of the reference 3D model. Furthermore, few 3D-QA metrics consider color information, which significantly restricts their effectiveness and scope of application. In this paper, we propose a no-reference (NR) quality assessment metric for colored 3D models represented by both point cloud and mesh. First, we project the 3D models from 3D space into quality-related geometry and color feature domains. Then, the 3D natural scene statistics (3D-NSS) and entropy are utilized to extract quality-aware features. Finally, machine learning is employed to regress the quality-aware features into visual quality scores. Our method is validated on the colored point cloud quality assessment database (SJTU-PCQA), the Waterloo point cloud assessment database (WPC), and the colored mesh quality assessment database (CMDM). The experimental results show that the proposed method outperforms most compared NR 3D-QA metrics with competitive computational resources and greatly reduces the performance gap with the state-of-the-art FR 3D-QA metrics. The code of the proposed model is publicly available now to facilitate further research.

Confidentiality in medical images through a genetic-based steganography algorithm in artificial intelligence.

Nowadays, image steganography has an important role in hiding information in advanced applications, such as medical image communication, confidential communication and secret data storing, protection of data alteration, access control system for digital content distribution and media database systems. In these applications, one of the most important aspects is to hide information in a cover image whithout suffering any alteration. Currently, all existing approaches used to hide a secret message in a cover image produce some level of distortion in this image. Although these levels of distortion present acceptable PSNR values, this causes minimal visual degradation that can be detected by steganalysis techniques. In this work, we propose a steganographic method based on a genetic algorithm to improve the PSNR level reduction. To achieve this aim, the proposed algorithm requires a private key composed of two values. The first value serves as a seed to generate the random values required on the genetic algorithm, and the second value represents the sequence of bit locations of the secret medical image within the cover image. At least the seed must be shared by a secure communication channel. The results demonstrate that the proposed method exhibits higher capacity in terms of PNSR level compared with existing works.

Deep learning based system for garment visual degradation prediction for longevity

Prolonging garment longevity is a well-recognized key strategy to reduce the overall environmental impact in the textile and clothing sector. In this context, change or degradation in esthetic or visual appeal of a garment with usage is an important factor that largely influence its longevity. Therefore, to engineer the garments for a required lifetime or prolong longevity, there is a need for predictive systems that can forecast the trajectory of visual degradation based on material/structural parameters or use conditions that can guide the practitioners for an optimal design. This paper develops a deep learning based predictive system for washing-induced visual change or degradation of selected garment areas. The study follows a systematic experimental design to generate and capture visual degradation in garment and equivalent fabric samples through 70 cycles in a controlled environment following guideline from relevant washing standards. Further, the generated data is utilized to train conditional Generative Adversarial Network-based deep learning model that learns the degradation pattern and links it to washing cycles and other seam properties. In addition, the predicted results are compared with experimental data using Frechet Inception Distance, to ascertain that the system prediction are visually similar to the experimental data and the prediction quality improves with training process.

Mid-life degradation evaluation of polycrystalline Si solar photovoltaic modules in a 100 kWp grid-tied system in east-central India

This study presents the mid-life degradation evaluation of a 100 kWp grid-tied solar photovoltaic (PV) power plant which is in the eleventh year of its operation. The field evaluation includes the visual inspections to identify the extent of degradation present in the PV modules, and the measurement of field current-voltage (I–V) data for maximum power deviation of the operational PV modules. Significant visual degradations were observed in many PV modules like encapsulant discoloration, snail tracks and cell cracks, burn marks on the surface of the PV module, shattering of front PV glass due to lightning strike, discoloration on metallization, corroded metal frame and hot spots. The PV module having heavy snail tracks, visible cracks and burn marks on the surface showed the highest power reduction of 95.9 W with respect to the reference module. The power degradation rate of the reference PV modules, which had no apparent visible degradation, over a period of ten years, was found to be varying between 0.16%/year to 1.72%/year. This study would be helpful in better understanding of the various types of degradations affecting the PV modules operating in the composite climate regions of east-central India.

Enlightening Low-Light Images With Dynamic Guidance for Context Enrichment

Images acquired in low-light conditions suffer from a series of visual quality degradations, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">e.g.</i> , low visibility, degraded contrast, and intensive noise. These complicated degradations based on various contexts ( <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">e.g</i> ., noise in smooth regions, over-exposure in well-exposed regions and low contrast around edges) cast major challenges to the low-light image enhancement. Herein, we propose a new methodology by imposing a learnable guidance map from the signal and deep priors, making the deep neural network adaptively enhance low-light images in a region-dependent manner. The enhancement capability of the learnable guidance map is further exploited with the multi-scale dilated context collaboration, leading to contextually enriched feature representations extracted by the model with various receptive fields. Through assimilating the intrinsic perceptual information from the learned guidance map, richer and more realistic textures are generated. Extensive experiments on real low-light images demonstrate the effectiveness of our method, which delivers superior results quantitatively and qualitatively. The code is available at <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><uri>https://github.com/lingyzhu0101/GEMSC</uri></i> to facilitate future research.

“Two-in-one” sulfur and nitrogen co-doped fluorescent silicon nanoparticles: Simultaneous as the fluorescent probe and photocatalyst for in-situ real time visual monitoring and degradation of tetracycline antibiotics

Detection and removal of contaminants are significant for environmental monitoring and remediation. In the present study, “two-in-one” silicon nanoparticles (SiNPs) were designed and prepared to simultaneously act as the fluorescent probe and degradation catalyst to detect and remove tetracycline (TCs) antibiotics. Thiourea and 3-aminopropyltrimethoxysilane were dopant and silicon source to generate fluorescent sulfur and nitrogen co-doped SiNPs (SN-SiNPs). The blue fluorescence of SN-SiNPs was selectively quenched by TCs due to the inner filter effect, whilst accompanied by the newly appeared yellow-green fluorescence resulting from aggregation induced fluorescence emission effect. Based on this phenomenon, SN-SiNPs can be used as fluorescent colorimetric probes for detection of doxycycline, oxytetracycline and tetracycline with limits of detection of 1.8 μg/L, 3.0 μg/L and 4.2 μg/L, respectively; the semi-quantitation can even be visually achieved by naked eyes. Particularly, SN-SiNPs were capable to catalyze the degradation of the three TCs effectively, achieving the removal rates of doxycycline, oxytetracycline and tetracycline of >90 %, >80 % and > 70 % after 240 min exposure to UV light. The catalytic ability of SN-SiNPs was derived from hydroxyl radical (•OH−), superoxide radical (•O2−) and singlet oxygen (1O2) produced by SN-SiNPs under UV irradiation. Moreover, integrating the fluorescent probe and photocatalyst together, the proposed SN-SiNPs simultaneously realized catalyzing the degradation of the three TCs and in-situ visually monitoring of the degradation process in real time. This study innovatively proposed an integrated probe for the detection and catalytic degradation of TCs, providing a new “two-in-one” strategy for rapid and simple detection and removal of drug pollutants.

One More Check: Making “Fake Background” Be Tracked Again

The one-shot multi-object tracking, which integrates object detection and ID embedding extraction into a unified network, has achieved groundbreaking results in recent years. However, current one-shot trackers solely rely on single-frame detections to predict candidate bounding boxes, which may be unreliable when facing disastrous visual degradation, e.g., motion blur, occlusions. Once a target bounding box is mistakenly classified as background by the detector, the temporal consistency of its corresponding tracklet will be no longer maintained. In this paper, we set out to restore the bounding boxes misclassified as ``fake background'' by proposing a re-check network. The re-check network innovatively expands the role of ID embedding from data association to motion forecasting by effectively propagating previous tracklets to the current frame with a small overhead. Note that the propagation results are yielded by an independent and efficient embedding search, preventing the model from over-relying on detection results. Eventually, it helps to reload the ``fake background'' and repair the broken tracklets. Building on a strong baseline CSTrack, we construct a new one-shot tracker and achieve favorable gains by 70.7 ➡ 76.4, 70.6 ➡ 76.3 MOTA on MOT16 and MOT17, respectively. It also reaches a new state-of-the-art MOTA and IDF1 performance. Code is released at https://github.com/JudasDie/SOTS.

Low complexity template-based watermarking with neural networks and various embedding templates

The popularity of watermarking was amplified due to the emergence of new immersive applications that connect the digital and physical worlds such as reading with a smartphone a watermark located in a physical object that leads to an online, dynamic source. Such applications require new watermarking schemes which must be robust enough to withstand significant watermark distortions and have low complexity in the sense that they must be adapted to be used on devices with constrained resources that are common in today's cyber-physical world. The key contribution of this work is the novel robust and low-complex template-based watermarking scheme that uses neural networks. Neural networks have been applied as they have already shown a distinguishable result in such image processing tasks as object detection and image restoration. The experiments show that our scheme outperforms the existing state-of-the-art template-based schemes in terms of robustness without visual degradation.

Deep Learning Enhanced UAV Imagery for Critical Infrastructure Protection

Unmanned aerial vehicles (UAVs) have seen a significant increase in their commercial application because of various technological break-throughs. As UAVs are typically used in open environments for purposes such as military applications, surveillance, and delivery of commodities, they rely primarily on the visual signals recorded by the flying UAV. UAVs can be used for critical infrastructure protection where their surveillance capabilities are used for monitoring these sites. While carrying out such missions in open environments, visual degradation is an unavoidable concern. It has a negative impact on the performance and security of the system. We propose a deep-learning-based framework, called Aerialgan, to solve the visual degradation caused by haze in the atmosphere. We trained the proposed model using an adversarial training algorithm which is commonly known as the generative adversarial ntwork (GAN) and aims to enhance the hazy images collected by a UAV and generate a clean, haze-free image of the same scene. In addition, we present the Aerial Non-Homogeneous Hazy (ANHH) dataset, which contains over 66,000 pairs of hazy and ground truth aerial photos with realistic, non-homogeneous haze of various densities. We used performance metrics such as peak signal-to-noise ratio and structural similarity index to evaluate our model on ANHH and compare it with contemporary state-of-the-art techniques in image dehazing. The proposed technique can be very useful in improving the reliability of surveillance where UAVs are used for critical infrastructure protection applications.

Rewritable two-dimensional DNA-based data storage with machine learning reconstruction

DNA-based data storage platforms traditionally encode information only in the nucleotide sequence of the molecule. Here we report on a two-dimensional molecular data storage system that records information in both the sequence and the backbone structure of DNA and performs nontrivial joint data encoding, decoding and processing. Our 2DDNA method efficiently stores images in synthetic DNA and embeds pertinent metadata as nicks in the DNA backbone. To avoid costly worst-case redundancy for correcting sequencing/rewriting errors and to mitigate issues associated with mismatched decoding parameters, we develop machine learning techniques for automatic discoloration detection and image inpainting. The 2DDNA platform is experimentally tested by reconstructing a library of images with undetectable or small visual degradation after readout processing, and by erasing and rewriting copyright metadata encoded in nicks. Our results demonstrate that DNA can serve both as a write-once and rewritable memory for heterogenous data and that data can be erased in a permanent, privacy-preserving manner. Moreover, the storage system can be made robust to degrading channel qualities while avoiding global error-correction redundancy.

Does Changing the Font Type Affect the Processing of Words Written in Cyrillic and Latin Alphabet?

To understand the reading process, it is necessary to explore the mechanisms of visual word recognition. The basic level of that recognition is the processing of letters, their size and visual identity. The specificity of the Serbian language is characterized by the parallel use of two alphabetic systems – Latin and Cyrillic, which contain a number of common, but also their own unique letters. Since some of the differences between the fonts are based on visual specific aditions at the end of letter's lines, there is also a significant contribution of fonts in letter recognition and reading. The main goal of this study is to examine the effect of font type on the processing of Latin and Cyrillic words. The aim was also to examine the effect of letter degradation on the word processing in these two alphabetic systems. The study included two experiments with Latin and Cyrillic words written in lowercase and uppercase. Participants were 221 students from the University of Banja Luka. Three factors were varied in both experiments: alphabet, “visual availability” (which refers to the visibility of words after degradation), and font type. Two analyses were performed ANOVA by subject (F1 analysis) and ANOVA by item (F2 analysis). Results show that there is a main effect of visual availability on processing speed of words. Visual degradation has slowed reaction time, but this effect is not the same in Latin and Cyrillic words. Significant interaction of font and alphabet is confirmed only in F2 analysis, so these results have limited validity. This study also revealed differences between lowercase and uppercase. Degradation of lowercase was more detrimental that degradation of uppercase. The obtained results indicate that word processing in two alphabetic systems can be partly explained by the visual characteristics and grapheme structure of their letters.

Two‐stage single image dehazing network using swin‐transformer

Hazy images often have color distortion, blur and other visible visual quality degradation, affecting the performance of some advanced visual tasks. Therefore, single image dehazing has always been a challenging and significant problem. Convolutional neural network has been widely used in image dehazing task, but the limitations of convolutional operation limit the development of dehazing task. Nowadays, Transformer offers a holistic approach to CV development and does not grow in location as the network deepens. For this reason, a hierarchical Transformer is introduced for use in the dehazing network. Specifically, the codec is improved and Transformer and CNN are combined to achieve basic feature extraction in the first stage. The encoder only models the global relationship at each layer, reducing the resolution of the feature map continuously and expanding the field of perception. In addition, an inter-block supervision mechanism is added between encoder unit and decoder unit to refine features and supervise and select them, thus improving the efficiency of feature transmission. In the second stage, the original resolution block is used to extract the local features, and then feature fusion and interaction are carried out. In addition, to ensure the authenticity of the transmission of characteristic signals in the first stage and improve the transmission efficiency of the network, fusion attention mechanism is added between stages. It adds the residual image of the early input features to the image acquired in the first stage, then passes to the next stage. Ablation experiments show that the two-stage network has significant benefits for image quality and visual effects. The experimental results on RESIDE, O-Haze, and I-Haze datasets show that the method is superior to advanced methods in dehazing effectiveness.

Comparative Genomics Provides Insights into Adaptive Evolution in Tactile-Foraging Birds.

Tactile-foraging birds have evolved an enlarged principal sensory nucleus (PrV) but smaller brain regions related to the visual system, which reflects the difference in sensory dependence. The “trade-off” may exist between different senses in tactile foragers, as well as between corresponding sensory-processing areas in the brain. We explored the mechanism underlying the adaptive evolution of sensory systems in three tactile foragers (kiwi, mallard, and crested ibis). The results showed that olfaction-related genes in kiwi and mallard and hearing-related genes in crested ibis were expanded, indicating they may also have sensitive olfaction or hearing, respectively. However, some genes required for visual development were positively selected or had convergent amino acid substitutions in all three tactile branches, and it seems to show the possibility of visual degradation. In addition, we may provide a new visual-degradation candidate gene PDLIM1 who suffered dense convergent amino acid substitutions within the ZM domain. At last, two genes responsible for regulating the proliferation and differentiation of neuronal progenitor cells may play roles in determining the relative sizes of sensory areas in brain. This exploration offers insight into the relationship between specialized tactile-forging behavior and the evolution of sensory abilities and brain structures.

Attentional control and metacognitive monitoring of the effects of different types of task-irrelevant sound on serial recall.

The presence of task-irrelevant sound disrupts short-term memory for serial information. Recent studies found that enhanced perceptual task-encoding load (static visual noise added to target items) reduces the disruptive effect of an auditory deviant but does not affect the task-specific interference by changing-state sound, indicating that the deviation effect may be more susceptible to attentional control. This study aimed to further specify the role of attentional control in shielding against different types of auditory distraction, examining speech and nonspeech distractors presented in laboratory and Web based experiments. To further elucidate the role of controlled processes, we tested whether the detrimental effects of distractor sounds-and their modulation by attentional control-reach participants' awareness. We found that changing-state sound and auditory deviants in steady-state sound equally affected both objective recall performance and metacognitive confidence judgments but did not affect the accuracy of confidence judgments. Most importantly, across four experiments, an increase of task load (visual degradation of the to-be-remembered items) did not reduce either type of auditory distraction. A close replication of the original modulation of the deviation effect by perceptual task load (in an online environment) even revealed a stronger deviation effect at high task load, suggesting that the manipulation may have influenced cognitive load and the ability to control distractor interference in memory. In line with a unitary account of auditory distraction, the results suggest that although both types of distraction reach metacognitive awareness, they may be equally unrelated to perceptual load and the availability of attentional resources. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

A Deep Learning Scheme for Efficient Multimedia Iot Data Compression

Given the voluminous nature of the multimedia sensed data, the Multimedia Internet of Things (MIoT) devices and networks will present several limitations in terms of power and communication overhead. One traditional solution to cope with the large-size data challenge is to use lossy compression. However, current lossy compression schemes require low compression rate to guarantee acceptable perceived image quality, which results in a low reduction of the communicated data size and consequently a low reduction in the energy and bandwidth consumption. Thus, an efficient compression solution is required for striking a good balance between data size (and consequently communication overhead) and visual degradation. In this paper, a Deep-Learning (DL) super-resolution model is applied to recuperate high quality images (at the application server side) given as input degraded images with a high compression ratio (at the sender side). The experimental analysis shows the effectiveness of the proposed solution in enhancing the visual quality of the compressed and down-scaled images. Consequently, the proposed solution reduces the overall communication overhead and power consumption of limited MIoT devices.

Blind Tone-Mapped HDR Image Quality Measurement by Analysis of Low-Level and High-Level Perceptual Characteristics

Tone-mapping technology aims to transform High Dynamic Range (HDR) images into standard dynamic range images for visualizing on standard screens. However, various visual artifacts will inevitably be induced during the tone-mapping process. To quantify the visual quality degradation for Tone-Mapped Images (TMIs) accurately and automatically, a robust blind TMI quality measurement method is developed by analysis of low-level and high-level perceptual characteristics. Specifically, for low-level visual features, considering that the tone-mapping operator is prone to destroy the local image contrast, which conveys vital structural information, we resort to the joint statistical features based on gradient map and Laplacian of Gaussian response to portray the local image contrast variation in TMIs. In addition, Chromatic Local Binary Pattern (CLBP) is leveraged to measure the colorfulness degradation over four chromatic descriptor maps. For high-level quality-aware features, deep-learned features are utilized by a pre-trained convolutional neural network to characterize the semantics variation. Finally, the extracted low-level and high-level features are combined and mapped into an overall quality score by regression function. Extensive experiments show that our metric achieves a higher performance over two publicly available TMI benchmark databases.

Intermediate-Layer Transferable Adversarial Attack With DNN Attention

The widespread deployment of deep learning models in practice necessitates an assessment of their vulnerability, particularly in security-sensitive areas. As a result, transfer-based adversarial attacks have elicited increasing interest in assessing the security of deep learning models. However, adversarial samples usually exhibit poor transferability over different models because of overfitting of the particular architecture and feature representation of a source model. To address this problem, the Intermediate Layer Attack with Attention guidance (IAA) is proposed to alleviate overfitting and enhance the black-box transferability. The IAA works on an intermediate layer <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">l</i> of the source model. Guided by the model’s attention to the features of layer <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">l</i> , the attack algorithm seeks and undermines the key features that are likely to be adopted by diverse architectures. Significantly, IAA focuses on improving existing white-box attacks without introducing significant visual perceptual quality degradation. Namely, IAA maintains the white-box attack performance of the original algorithm while significantly enhancing its black-box transferability. Extensive experiments on ImageNet classifiers confirmed the effectiveness of our method. The proposed IAA outperformed all state-of-the-art benchmarks in various white-box and black-box settings.

The transpiration and respiration as mechanisms of water loss in cold storage of figs

Transpiration and respiration are two mechanisms of water loss in fresh agricultural products, resulting in visual and texture degradation. Neglecting respiration as a mechanism of water loss may lead to erroneous results at saturation where water vapour pressure deficit is zero and thus water loss is expected to be zero, however, the existence of a finite water loss is noted. In this context, an analysis of the associated with transpiration and respiration water loss in figs (Ficus carica L.) was carried out at 0oC, 10oC and 20oC and 45.64%, 80.22% and 98.65% relative humidity as well as the air conditions of walk-in cold storage rooms. The estimated transpiration rate ranged between 0.11-1.416 mg cm-2 h -1 for a water vapour pressure deficit of 0.0-0.98 kPa. The water vapour pressure deficit estimation was based on the difference between cold air temperature and figs’ surface temperature. The respiration rate was calculated at 0oC, 10oC and 20oC as 0.47±0.08, 0.94±0.11 and 2.69±0.17 mLCO2100g-1 h -1 . Quantification of the water loss showed that at 20oC and saturation, the water loss due to respiration accounts for 3.9% of the respective water loss due to water vapour pressure deficit while on average, the water loss due to respiration accounts for 1.5%, 2.1% and 2.6% of the water loss due to water vapour pressure deficit at 0oC, 10oC and 20oC.