Naked Human Eye Research Articles

An approach to automatic fault detection in four-point system for knitted fabric with our benchmark dataset Isl-Knit

Knit fabric is one of the dominating fabric types for wearable textile across the whole world and during the production of knit fabric, faults created by different reasons cause difficulties in the subsequent process. The existing fault detection process in the knitting industry is done manually by the human naked eye. To detect faults automatically, deep learning-based models are very efficient and can reduce the workload for fabric inspection. However, implementing a deep learning-based model for fabric fault detection needs a large number of images with different types of knit fabric faults from real-world scenarios to train and for better performance. Most of the existing fabric fault datasets are based on woven fabric, which cannot be used for training because the faults of woven fabric and knit fabrics are completely different. In this research, we propose a new dataset named ISL-Knit. It is a benchmark dataset for knit fabric faults collected from different knit dyeing industries in Bangladesh. Our dataset contains high-resolution images of grey fabric and dyed fabric annotated with 7 types of faults. To provide correct annotation, all images of faults are annotated and verified with the proper reference which can be a great tool for boosting up deep learning-based models for automatic knit fabric fault detection in the textile field. A comparison of the accuracy of the deep learning-based model by training with our proposed dataset and existing knit fabric fault dataset is also done. To develop an automatic fault detection system, we trained the YOLOv5 models with our dataset, considering different image sizes. For real-world implementation, the best models are selected considering mAP and inference time. YOLOv5 models are implemented in Raspberry Pi with a Raspberry Pi camera, a laptop with no dedicated GPU, and a laptop with a dedicated GPU with a web camera to observe the performance. Finally, an automated four-point inspection system is developed by calibrating the camera to generate the score representing the quality of fabrics. The implementation represents the feasibility of a deep learning-based model in knit fabric fault detection, considering the cost and accuracy.

Changes in Metabolism and Content of Chlorophyll in Common Duckweed (Lemna minor L.) Caused by Environmental Contamination with Fluorides.

The impact of fluorine on plants remains poorly understood. We examined duckweed growth in extracts of soil contaminated with fluorine leached from chicken manure. Additionally, fluorine levels were analyzed in fresh manure, outdoor-stored manure, and soil samples at varying distances from the manure pile. Fresh manure contained 37-48 mg F- × kg-1, while soil extracts contained 2.1 to 4.9 mg F- × kg-1. We evaluated the physiological effects of fluorine on duckweed cultured on soil extracts or in 50% Murashige-Skoog (MS) medium supplemented with fluorine concentrations matching those in soil samples (2.1 to 4.9 mg F- × L-1), as well as at 0, 4, and 210 mg × L-1. Duckweed exposed to fluorine displayed similar toxicity symptoms whether in soil extracts or supplemented medium. Fluoride at concentrations of 2.1 to 4.9 mg F- × L-1 reduced the intact chlorophyll content, binding the porphyrin ring at position 32 without affecting Mg2+. This reaction resulted in chlorophyll a absorption peak shifted towards shorter wavelengths and formation of a new band of the F--chlorophyll a complex at λ = 421 nm. Moreover, plants exposed to low concentrations of fluorine exhibited increased activities of aminolevulinic acid dehydratase and chlorophyllase, whereas the activities of both enzymes sharply declined when the fluoride concentration exceeded 4.9 mg × L-1. Consequently, fluorine damages chlorophyll a, disrupts the activity of chlorophyll-metabolizing enzymes, and diminishes the plant growth rate, even when the effects of these disruptions are too subtle to be discerned by the naked human eye.

MRI radiomics captures early treatment response in patient-derived organoid endometrial cancer mouse models.

Radiomics can capture microscale information in medical images beyond what is visible to the naked human eye. Using a clinically relevant mouse model for endometrial cancer, the objective of this study was to develop and validate a radiomic signature (RS) predicting response to standard chemotherapy. Mice orthotopically implanted with a patient-derived grade 3 endometrioid endometrial cancer organoid model (O-PDX) were allocated to chemotherapy (combined paclitaxel/carboplatin, n=11) or saline/control (n=13). During tumor progression, the mice underwent weekly T2-weighted (T2w) magnetic resonance imaging (MRI). Segmentation of primary tumor volume (vMRI) allowed extraction of radiomic features from whole-volume tumor masks. A radiomic model for predicting treatment response was derived employing least absolute shrinkage and selection operator (LASSO) statistics at endpoint images in the orthotopic O-PDX (RS_O), and subsequently applied on the earlier study timepoints (RS_O at baseline, and week 1-3). For external validation, the radiomic model was tested in a separate T2w-MRI dataset on segmented whole-volume subcutaneous tumors (RS_S) from the same O-PDX model, imaged at three timepoints (baseline, day 3 and day 10/endpoint) after start of chemotherapy (n=8 tumors) or saline/control (n=8 tumors). The RS_O yielded rapidly increasing area under the receiver operating characteristic (ROC) curves (AUCs) for predicting treatment response from baseline until endpoint; AUC=0.38 (baseline); 0.80 (week 1), 0.85 (week 2), 0.96 (week 3) and 1.0 (endpoint). In comparison, vMRI yielded AUCs of 0.37 (baseline); 0.69 (w1); 0.83 (week 2); 0.92 (week 3) and 0.97 (endpoint). When tested in the external validation dataset, RS_S yielded high accuracy for predicting treatment response at day10/endpoint (AUC=0.85) and tended to yield higher AUC than vMRI (AUC=0.78, p=0.18). Neither RS_S nor vMRI predicted response at day 3 in the external validation set (AUC=0.56 for both). We have developed and validated a radiomic signature that was able to capture chemotherapeutic treatment response both in an O-PDX and in a subcutaneous endometrial cancer mouse model. This study supports the promising role of preclinical imaging including radiomic tumor profiling to assess early treatment response in endometrial cancer models.

Effect of various beverage immersions on the color stability and degree of cure of two types of dental resin composite materials: An in vitro study

Among the wide range of polymer matrices, polymethyl methacrylate (PMMA)-based matrices can be considered the most common polymer in dental applications. Significant parameters, such as conversion degree and change of color of two commercial brands, nano-filled composite resin, NovaPro (N), and micro-hybrid composite resin, VladMiVa (V), were investigated after curing and immersing the dental specimens in different drinks for one and seven days at 37°C. Raman spectroscopy was used to evaluate the degree of conversion (DC) by measuring the ratio of the absorption bands related to aromatic and methacrylate C=C stretching modes. These ratios were of the same value. CIE (Color Imaging Experiment) was used to investigate the color change of the filling resins. The N composite has a noticeable color change in coffee after 24 hours and almost the exact change after one week. Yogurt also significantly affects the color stability of the same composite just after one week of immersion. Cola dramatically affects the color of the same composite after one week of immersion. For the (V) composite, the effect of the drinks after immersion for 24 hours and after one week was within the acceptable limit and non-noticeable by the naked human eye, which means it has better color stability than the (N).

Radiomics and artificial intelligence for risk stratification of pulmonary nodules: Ready for primetime?

Pulmonary nodules are ubiquitously found on computed tomography (CT) imaging either incidentally or via lung cancer screening and require careful diagnostic evaluation and management to both diagnose malignancy when present and avoid unnecessary biopsy of benign lesions. To engage in this complex decision-making, clinicians must first risk stratify pulmonary nodules to determine what the best course of action should be. Recent developments in imaging technology, computer processing power, and artificial intelligence algorithms have yielded radiomics-based computer-aided diagnosis tools that use CT imaging data including features invisible to the naked human eye to predict pulmonary nodule malignancy risk and are designed to be used as a supplement to routine clinical risk assessment. These tools vary widely in their algorithm construction, internal and external validation populations, intended-use populations, and commercial availability. While several clinical validation studies have been published, robust clinical utility and clinical effectiveness data are not yet currently available. However, there is reason for optimism as ongoing and future studies aim to target this knowledge gap, in the hopes of improving the diagnostic process for patients with pulmonary nodules.

A deep learning model for rapid classification of tea coal disease

BackgroundThe common tea tree disease known as “tea coal disease” (Neocapnodium theae Hara) can have a negative impact on tea yield and quality. The majority of conventional approaches for identifying tea coal disease rely on observation with the human naked eye, which is labor- and time-intensive and frequently influenced by subjective factors. The present study developed a deep learning model based on RGB and hyperspectral images for tea coal disease rapid classification.ResultsBoth RGB and hyperspectral could be used for classifying tea coal disease. The accuracy of the classification models established by RGB imaging using ResNet18, VGG16, AlexNet, WT-ResNet18, WT-VGG16, and WT-AlexNet was 60%, 58%, 52%, 70%, 64%, and 57%, respectively, and the optimal classification model for RGB was the WT-ResNet18. The accuracy of the classification models established by hyperspectral imaging using UVE-LSTM, CARS-LSTM, NONE-LSTM, UVE-SVM, CARS-SVM, and NONE-SVM was 80%, 95%, 90%, 61%, 77%, and 65%, respectively, and the optimal classification model for hyperspectral was the CARS-LSTM, which was superior to the model based on RGB imaging.ConclusionsThis study revealed the classification potential of tea coal disease based on RGB and hyperspectral imaging, which can provide an accurate, non-destructive, and efficient classification method for monitoring tea coal disease.

A Study on Arūpā Brahma in Theravāda Buddhism

While writing this topic of ‘A Study on Arūpā Brahma in Theravāda Buddhism”, about seven articles written on this type of Brahmas, some Samahta papers and Ahidhamma books were reviewed, studied and analyzed. Digha Nikaya Sutta translation (III), Parhtika web was also studied and reviewed. This paper points out that there are very few papers written on this topic and we can only get some kind of knowledge on Arūpā Brahma only from “DIALOGUES OF THE BUDDHA” articles and some papers written by our famous Buddhist Sayadaws. However, after reading this paper, one will hope to become believed in that : there are Arūpā Brahma realms in this Universe which we cannot see with our human naked eyes and these Arūpā Brahma realms are unsuitable places for beings to be reborn because these realms are included in the eight untimely, unseasonable intervals for life in a religious order.

A Green-Emitting Luminol Analogue as the Next-Generation Chemiluminescent Substrate in Biochemical Analysis.

Luminol and its derivatives are extensively used as chemiluminogenic substrates in bioimaging and biochemical analysis. Luminol reagents can typically emit blue chemiluminescence (CL), whose wavelength is normally outside the most sensitive detection range of human naked eyes and most CL analyzers with silicon-based charge-coupled device (CCD) detectors. Development of luminol analogues with longer wavelength emission is thus attractive. Herein, four new phthalhydrazide CL probes (GL-1/2/3/4) have been prepared through the derivatization of luminol. The most promising one, 5-(4-hydroxy-1,3-dioxoisoindolin-2-yl)-2,3-dihydrophthalazine-1,4-dione (GL-1), emits bright green CL upon oxidation and shows enhanced CL performance compared to its parent luminol. Bloodstain imaging, horseradish peroxidase (HRP)-based immunoassay, and the analysis of glucose/glucose oxidase reaction have been performed using the GL-1 reagent. These results indicate that GL-1 is a new chemiluminogenic luminol analogue with great potential in real analytical applications and will be an alternative to replace luminol in practical CL analysis.

Sustainable Utilization of Machine-Vision-Technique-Based Algorithm in Objective Evaluation of Confocal Microscope Images

Confocal microscope imaging has become popular in biotechnology labs. Confocal imaging technology utilizes fluorescence optics, where laser light is focused onto a specific spot at a defined depth in the sample. A considerable number of images are produced regularly during the process of research. These images require methods of unbiased quantification to have meaningful analyses. Increasing efforts to tie reimbursement to outcomes will likely increase the need for objective data in analyzing confocal microscope images in the coming years. Utilizing visual quantification methods to quantify confocal images with naked human eyes is an essential but often underreported outcome measure due to the time required for manual counting and estimation. The current method (visual quantification methods) of image quantification is time-consuming and cumbersome, and manual measurement is imprecise because of the natural differences among human eyes’ abilities. Subsequently, objective outcome evaluation can obviate the drawbacks of the current methods and facilitate recording for documenting function and research purposes. To achieve a fast and valuable objective estimation of fluorescence in each image, an algorithm was designed based on machine vision techniques to extract the targeted objects in images that resulted from confocal images and then estimate the covered area to produce a percentage value similar to the outcome of the current method and is predicted to contribute to sustainable biotechnology image analyses by reducing time and labor consumption. The results show strong evidence that t-designed objective algorithm evaluations can replace the current method of manual and visual quantification methods to the extent that the Intraclass Correlation Coefficient (ICC) is 0.9.

Democratizing Pathological Image Segmentation with Lay Annotators via Molecular-empowered Learning.

Multi-class cell segmentation in high-resolution Giga-pixel whole slide images (WSI) is critical for various clinical applications. Training such an AI model typically requires labor-intensive pixel-wise manual annotation from experienced domain experts (e.g., pathologists). Moreover, such annotation is error-prone when differentiating fine-grained cell types (e.g., podocyte and mesangial cells) via the naked human eye. In this study, we assess the feasibility of democratizing pathological AI deployment by only using lay annotators (annotators without medical domain knowledge). The contribution of this paper is threefold: (1) We proposed a molecular-empowered learning scheme for multi-class cell segmentation using partial labels from lay annotators; (2) The proposed method integrated Giga-pixel level molecular-morphology cross-modality registration, molecular-informed annotation, and molecular-oriented segmentation model, so as to achieve significantly superior performance via 3 lay annotators as compared with 2 experienced pathologists; (3) A deep corrective learning (learning with imperfect label) method is proposed to further improve the segmentation performance using partially annotated noisy data. From the experimental results, our learning method achieved F1 = 0.8496 using molecular-informed annotations from lay annotators, which is better than conventional morphology-based annotations (F1 = 0.7015) from experienced pathologists. Our method democratizes the development of a pathological segmentation deep model to the lay annotator level, which consequently scales up the learning process similar to a non-medical computer vision task. The official implementation and cell annotations are publicly available at https://github.com/hrlblab/MolecularEL.

Nails Deformation Sensing For Disease Detection Using Fusion Method For Texture Analysis

Medical image processing is essential for identifying conditions like skin cancer, brain tumors, breast cancer, and illnesses of the heart and lungs. To detect disorders, the human nail serves as a window into the body. Various methods are available in the healthcare industry for early disease diagnosis. One method for diagnosing diseases in their early stages is nail image analysis. It is possible to forecast some general and dermatological disorders using human nails. The naked human eye cannot detect numerous variations in nail color or texture. In this research, models for illness prediction using nail pictures are constructed. In this work, the fusion of statistical and structural texture featuresis employed to forecast diseases. Merging structural and statistical features allows for the analysis of the correctness of many models and the selection of the most relevant one. Numerous methods, particularly GLCM, LBP, and Gabor filter, were used to combine statistical and structural features

Effects of supplementing green light to red and blue light on the growth and yield of lettuce in plant factories

Red and blue light-emitting diodes (LEDs) are widely used as light sources in plant factories. Plants grown under red and blue light only produces purplish-gray light environment where plant leaves look purplish-gray. Under that light condition, it makes difficulty to check health status of plants specially insect and disease infected leaves by human naked eye. But the leaf color turns green when inspected under full spectrum of light environment. In this regard, the use of green light to red and blue create a white light environment which is congenial for the grower. Likewise, CO2 assimilation of green light is usually lower compared to red and blue light because of its lower absorptance under low PPFD. But at higher PPFD (≥500 µmol m−2 s−1), green light shows higher CO2 assimilation rate compared to red and blue light through uniform distribution of full spectrum (red, green and blue) of light into the plant canopy and the lower leaves. In this study, optimal intensity of green light supplementation to the red and blue LEDs was evaluated for better growth, yield, and quality of lettuce. In experiment I and II, green light was added with red and blue light where total photosynthetic photon flux density (PPFD) was increased for the additional green light. In experiment III, 0, 10, 20, 30, 40, 50, 60, 70, 80 µmol m−² s−1 green (G) light was supplemented to 235 µmol m−² s−1 red (R) and 59 µmol m−² s−1 blue (B) light maintaining 294 µmol m−² s−1 total PPFD. Lettuce plants were cultivated hydroponically in three-step vertical grow beds using half-strength of Enshi nutrient solution. The temperature was maintained at 20 ± 2 °C at day and night and the day length was 16 h. In experiment I and II, the fresh mass of lettuce grown under the combination of R, B and G LEDs was found higher than those cultivated under R and B LEDs only. In experiment III, lettuce plants produced high fresh masses when 30 µmol m−² s−1 of G light was supplemented with R and B lights maintaining the ratio 211:30:53 (R:G:B). By supplementing excessive G light (≥50 µmol m−² s−1) with R and B lights, lettuce shoot fresh mass declined. The various combinations of irradiation of R, G and B LED did not significantly affect the contents of ascorbic acid and mineral nutrients. It is recommended to supplement 30 µmol m−2 s−1 of green light to red and blue light for a higher yield of lettuce. Therefore, the suitable LED (R:G:B) combination would be 72% of red, 10% of green and 18% of blue for lettuce cultivation following recycled hydroponics in plant factories.

FRACTALS. ORIGINS AND HISTORY

Fractal shapes can be found everywhere around us in Nature, at small scale microscopic level, at normal scale observable with the naked human eye, but also at very large scale level, when looking at very far distant galaxies. As it will be presented in this paper, there is an inherent relationship between fractal structures in Nature and the harmony perceived by human eye in the structures of Nature. Although not explicitly formulated until late XII century by Fibonacci, the very same Golden Proportion had been already approached by the ancient Greeks. This Divine Proportion which can be found around us in Nature, represents the base for the latter concept of fractal and is also the link between early mathematics and advanced fractal mathematics. Some of the first ideas about harmony and proportions are dating back in Antiquity and they were promoted by the ancient Greeks in their philosophical schools, by scholars like Euclid, Thales and Pitagora. Although the concept of fractal it is a rather young one, being introduced by Mandelbrot in the ‘70s of the XX th century, the origins of the concept of fractal seem to be found, much earlier, in the aforementioned philosophical schools. In their philosophical schools, the antics, as far as we know, beginning with the Greeks, had discovered and used the concepts of proportion and harmony, in certain proportions they worked with. One such proportion is represented by the Golden Ratio.

Spotting micro‐movements in image sequence by introducing intelligent cubic‐LBP

Invisible to naked human eyes, micro-movements are barely noticeable. There are wide ranges of micro-movement applications from spotting subtle changes in a volcano, vascular pulse, and blood vessel to micro-expression detection. In the latter, some evil thoughts can be unveiled, resulting in identifying crooks and lawbreakers and it arises from involuntary subtle and short-duration of facial muscles movements. Precise spotting of these tiny movements is possible only when multiple aspects of temporal images are scrutinized. Meanwhile, since motions often happen in one or two directions, it is rudimentary to extract complete feature sets in textural-based approaches such as cubic Local Binary Pattern (cubic-LBP). Approaches like cubic-LBP also have imposed an unnecessary computation burden. Hence, in this research, a novel method named intelligent cubic-LBP is proposed, which incorporates Convolutional Neural Network (CNN) model. This model learns to select the useful plane(s) automatically. Apex is then detected by applying Partial Autocorrelation Coefficient (PACF) on the selected plane(s). The experimental results show significant improvement in micro-movements identification. The accuracy of the apex frame detection has elevated to 10% and 17% in the Chinese Academy of Sciences Micro-Expressions (CASME) and the CASME II databases, respectively.

A Random Forest-based Classifier for MYCN Status Prediction in Neuroblastoma using CT Images.

Neuroblastoma (NB) is the most common extracranial solid tumor in childhood. Genomic amplification of MYCN is associated with poor outcomes and is detected in 16% of all NB cases. CT scans and MRI are the imaging techniques recommended for diagnosis and disease staging. The assessment of imaging features such as tumor volume, shape, and local extension represent relevant prognostic information. Radiogenomics have shown powerful results in the assessment of the genotype based on imaging findings automatically extracted from medical images. In this work, random forest was used to classify the MYCN amplification using radiomic features extracted from CT slices in a population of 46 NB patients. The learning model showed an area under the curve (AUC) of 0.85 ± 0.13, suggesting that radiomic-based methodologies might be helpful in the extraction of information that is not accessible by human naked eyes but could aid the clinicians on the diagnosis and treatment plan definition. Clinical relevance - This approach represents a random forest-based model to predict the MYCN amplification in NB patients that could give a faster, earlier, and repeatable analysis of the tumor along the time.

Color changes in bulk-fill resin composites as a result of visible light-curing.

We evaluated twelve commercially available bulk-fill resin composites to determine the colorimetric changes that occurred as a result of the photo-polymerization reaction. Samples of the resin composites were prepared as disks 8 mm in diameter and 3 mm in thickness. A colorimeter was used to measure the color of samples on a black background before and after the polymerization processed had been initiated. As results, each of the photo-initiated bulk-fill resin composites tested except for Beautifil Bulk A showed a significant color difference more than that of conventional resin composites. Bulk Base Flow became darker, whereas Beautifil Bulk Flow became clearly lighter. All resin composites except for Bulk Base Flow became less yellow. Thus, Beautifil Bulk A exhibited the least color change among all the bulk-fill resin composites. However, this color change showed a marked change (NBS units) as observed by the human naked eye.

Multi-Label Chest X-Ray Classification via Deep Learning

In this era of pandemic, the future of healthcare industry has never been more exciting. Artificial intelligence and machine learning (AI & ML) present opportunities to develop solutions that cater for very specific needs within the industry. Deep learning in healthcare had become incredibly powerful for supporting clinics and in transforming patient care in general. Deep learning is increasingly being applied for the detection of clinically important features in the images beyond what can be perceived by the naked human eye. Chest X-ray images are one of the most common clinical method for diagnosing a number of diseases such as pneumonia, lung cancer and many other abnormalities like lesions and fractures. Proper diagnosis of a disease from X-ray images is often challenging task for even expert radiologists and there is a growing need for computerized support systems due to the large amount of information encoded in X-Ray images. The goal of this paper is to develop a lightweight solution to detect 14 different chest conditions from an X ray image. Given an X-ray image as input, our classifier outputs a label vector indicating which of 14 disease classes does the image fall into. Along with the image features, we are also going to use non-image features available in the data such as X-ray view type, age, gender etc. The original study conducted Stanford ML Group is our base line. Original study focuses on predicting 5 diseases. Our aim is to improve upon previous work, expand prediction to 14 diseases and provide insight for future chest radiography research.

A smart agriculture framework for IoT based plant decay detection using smart croft algorithm

Internet of things (IOT) is an assuring technology which provides systematic and logical solutions towards the revolution of various realms. Various researches and analysis have been conducted and numerous methods have been incorporated to apply IoT technology on agronomical fields. IoT can play a vital role in timely detection of declining plant health so that appropriate measures can be taken. It is a huge step towards smart agriculture. In this paper we propose a model to build up an automated framework which will recognize the crop decay in the initial phase which is imperceptible to naked human eyes. This model helps in prevention of huge losses and also save a lot of time and labor. The proposed model builds a recognition framework using sensors like humidity, moisture, temperature and color of the plant leaf. The data from the sensors is sent to Arduino to Cloud which then analyzes the data and helps in identifying the plant decay. In the upcoming years, the internet of things will be a crucial bit in the smart farming system.

Non-destructive visual inspection of photocurable coatings based on fluorescent response of naked-eye visible colorimetric and fluorescent sensors

• Terphenyl-based fluorescent sensors for monitoring photopolymerization processes are reported. • Colourimetric sensors for the investigation of cationic photopolymerization of epoxy monomer and vinyl monomer are developed. • Bimolecular photoinitiating systems have been successfully applied to 3D printing. • The changes in fluorescence colour before and after photopolymerization are possible using only the human naked-eye. Five novel 4,6-diphenyl-2-pyrrol-1-yl-benzene-1,3-dicarbonitrile derivatives bearing the following substituents: -H, -SCH3, –OCH3, –CN, -NPh2 were synthesized and investigated as fluorescent sensors for photopolymerization processes. It was turned out, that using these compounds provide a visible and direct strategy to monitor the photopolymerization process due to the change in fluorescence color during the photo-crosslinking process. Studied colorimetric sensors were investigated during free-radical and cationic photopolymerization as well as for the production of photocurable coatings. Furthermore, 4,6-diphenyl-2-pyrrol-1-yl-benzene-1,3-dicarbonitrile derivatives are also excellent candidates for photosensitizers of iodonium salt for photopolymerization processes, and thus they can play a double role, as a component of the photoinitiating system and sensor. Free radical of an acrylate monomer and cationic photopolymerization of vinyl and epoxy monomers was performed in the presence of two-component photoinitiating systems. Fluorescence spectroscopy was used to monitor fluorescence spectra changes during photopolymerization, what is more, it was also possible to direct observation fluorescence color changes by the naked human eye. Fourier transform infrared spectroscopy was applied for the determination of final conversion values after photopolymerization processes.

Do not get fooled: Defense against the one-pixel attack to protect IoT-enabled Deep Learning systems

With the rapid evolution and proliferation of Artificial Intelligence (AI) throughout the past decade, industry-wide Deep Learning (DL) applications within IIoT ecosystems are expected to rise exponentially within this coming decade. However, recent studies have highlighted the extreme sensitivity and vulnerability of modern Deep Neural Networks (DNNs) to adversarial examples. While being imperceptible and benign to the naked human eye, those adversarial perturbations crafted from one-pixel attacks can drastically impact the accuracy of DNNs. The consequences of those adversarial attacks in critical applications such as healthcare and autonomous vehicles can prove disastrous. Therefore, the urge to ensure the robustness and resilience of DL systems against one-pixel attacks has recently started attracting stupendous attention from both academicians and industry professionals. Throughout this paper, we first develop a one-pixel threat model within an IIoT ecosystem and propose a novel image recovery defense mechanism to detect and mitigate one-pixel attacks based on Accelerated Proximal Gradient approach. Experimental results proved that our proposed solution is highly effective and efficient in detecting and mitigating one-pixel attacks in state-of-the-art neural networks, more specifically LeNet and ResNet, using the CIFAR10 and MNIST datasets.