A Distributed Ant Colony Optimization in Edge Detection

With the rise of image data and increased complexity of tasks in edge detection, conventional artificial intelligence techniques have been severely impacted. To be able to solve even greaterproblems of the future, learning algorithms must maintain high speed and accuracy through economical means. Traditional edge detection approaches cannot detect edges in images in a timely manner due to memory and computational time constraints. In this work, a novel parallelized ant colony optimization technique in a distributed framework provided by the Hadoop/Map-Reduce infrastructure is proposed to improve the edge detection capabilities. Moreover, a filtering technique is applied to reduce the noisy background of images to achieve significant improvement in the accuracy of edge detection. Close examinations of the implementation of the proposed algorithm are discussed and demonstrated through experiments. Results reveal high classification accuracy and significant improvementsin speedup, scaleup and sizeup compared to the standard algorithms.

The Cascade Hilbert-Zero Decomposition: A Novel Method for Peaks Resolution and Its Application to Raman Spectra

Raman spectra of biological objects are sufficiently complex since they are comprised of wide diffusive spectral peaks over a noisy background. This makes the resolution of individual closely positioned components a complicated task. Here we propose a method for constructing an approximation of such systems by a series, respectively, to shifts of the Gaussian functions with different adjustable dispersions. It is based on the coordination of the Gaussian peaks’ location with the zeros of the signal’s Hilbert transform. The resolution of overlapping peaks is achieved by applying this procedure in a hierarchical cascade way, subsequently excluding peaks of each level of decomposition. Both the mathematical rationale for the localization of intervals, where the zero crossing of the Hilbert-transformed uni- and multimodal mixtures of Gaussians occurs, and the step-by-step outline of the numerical algorithm are provided and discussed. As a practical case study, we analyze results of the processing of a complicated Raman spectrum obtained from a strain of Mycobacterium tuberculosis. However, the proposed method can be applied to signals of different origins formed by overlapped localized pulses too.

Automated Neuron Tracing Using Content-Aware Adaptive Voxel Scooping on CNN Predicted Probability Map

Neuron tracing, as the essential step for neural circuit building and brain information flow analyzing, plays an important role in the understanding of brain organization and function. Though lots of methods have been proposed, automatic and accurate neuron tracing from optical images remains challenging. Current methods often had trouble in tracing the complex tree-like distorted structures and broken parts of neurite from a noisy background. To address these issues, we propose a method for accurate neuron tracing using content-aware adaptive voxel scooping on a convolutional neural network (CNN) predicted probability map. First, a 3D residual CNN was applied as preprocessing to predict the object probability and suppress high noise. Then, instead of tracing on the binary image produced by maximum classification, an adaptive voxel scooping method was presented for successive neurite tracing on the probability map, based on the internal content properties (distance, connectivity, and probability continuity along direction) of the neurite. Last, the neuron tree graph was built using the length first criterion. The proposed method was evaluated on the public BigNeuron datasets and fluorescence micro-optical sectioning tomography (fMOST) datasets and outperformed current state-of-art methods on images with neurites that had broken parts and complex structures. The high accuracy tracing proved the potential of the proposed method for neuron tracing on large-scale.

Fully-automated root image analysis (faRIA)

High-throughput root phenotyping in the soil became an indispensable quantitative tool for the assessment of effects of climatic factors and molecular perturbation on plant root morphology, development and function. To efficiently analyse a large amount of structurally complex soil-root images advanced methods for automated image segmentation are required. Due to often unavoidable overlap between the intensity of fore- and background regions simple thresholding methods are, generally, not suitable for the segmentation of root regions. Higher-level cognitive models such as convolutional neural networks (CNN) provide capabilities for segmenting roots from heterogeneous and noisy background structures, however, they require a representative set of manually segmented (ground truth) images. Here, we present a GUI-based tool for fully automated quantitative analysis of root images using a pre-trained CNN model, which relies on an extension of the U-Net architecture. The developed CNN framework was designed to efficiently segment root structures of different size, shape and optical contrast using low budget hardware systems. The CNN model was trained on a set of 6465 masks derived from 182 manually segmented near-infrared (NIR) maize root images. Our experimental results show that the proposed approach achieves a Dice coefficient of 0.87 and outperforms existing tools (e.g., SegRoot) with Dice coefficient of 0.67 by application not only to NIR but also to other imaging modalities and plant species such as barley and arabidopsis soil-root images from LED-rhizotron and UV imaging systems, respectively. In summary, the developed software framework enables users to efficiently analyse soil-root images in an automated manner (i.e. without manual interaction with data and/or parameter tuning) providing quantitative plant scientists with a powerful analytical tool.

Attention U-Net with Feature Fusion Module for Robust Defect Detection

U-Net shows a remarkable performance and makes significant progress for segmentation task in medical images. Despite the outstanding achievements, the common case of defect detection in industrial scenes is still a challenging task, due to the noisy background, unpredictable environment, varying shapes and sizes of the defects. Traditional U-Net may not be suitable for low-quality images with low illumination and corruption, which are often presented in the practical collections in real-world scenes. In this paper, we propose an attention U-Net with feature fusion module for combining multi-scale features to detect the defects in noisy images automatically. Feature fusion module contains convolution kernels of different scales to capture shallow layer features and combine them with the high-dimensional features. Meanwhile, attention gates are used to enhance the robustness of skip connection between the feature maps. The proposed method is evaluated on two datasets. The best precision rate and MIoU of defect detection are 95.6% and 92.5%. The best F-score of concrete crack detection is 95.0%. Experimental results show that the proposed approach achieves promising results in both datasets. It demonstrates that our approach consistently outperforms other U-Net-based approaches for defect detection in low-quality images. Experimental results have shown the possibility of developing a mixture system that can be deployed in many applications, such as remote sensing image analysis, earthquake disaster situation assessment, and so on.

A Pilot Experiment on Infrasonic Lahar Detection at Mount Adams, Cascades: Ambient Infrasound and Wind-Noise Characterization at a Quiescent Stratovolcano

Erosion, hydrothermal activity, and magmatism at volcanoes can cause large and unexpected mass wasting events. Large fluidized debris flows have occurred within the past 6000 yr at Mount Adams, Washington, and present a hazard to communities downstream. In August 2017, we began a pilot experiment to investigate the potential of infrasound arrays for detecting and tracking debris flows at Mount Adams. We deployed a telemetered four-element infrasound array (BEAR, 85 m aperture), ~11 km from a geologically unstable area where mass wasting has repeatedly originated. We present a preliminary analysis of BEAR data, representing a survey of the ambient infrasound and noise environment at this quiescent stratovolcano. Array processing reveals near continuous and persistent infrasound signals arriving from the direction of Mount Adams, which we hypothesize are fluvial sounds from the steep drainages on the southwest flank. We interpret observed fluctuations in the detectability of these signals as resulting from a combination of (1) wind-noise variations at the array, (2) changes in local infrasound propagation conditions associated with atmospheric boundary layer variability, and (3) changing water flow speeds and volumes in the channels due to freezing, thawing, and precipitation events. Suspected mass movement events during the study period are small (volumes <105 m3 and durations <2 min), with one of five visually confirmed events detected infrasonically at BEAR. We locate this small event, which satellite imagery suggests was a glacial avalanche, using three additional temporary arrays operating for five days in August 2018. Events large enough to threaten downstream communities would likely produce stronger infrasonic signals detectable at BEAR. In complement to recent literature demonstrating the potential for infrasonic detection of volcano mass movements (Allstadt et al., 2018), this study highlights the practical and computational challenges involved in identifying signals of interest in the expected noisy background environment of volcanic topography and drainages.

Separating Chinese Character from Noisy Background Using GAN

Separating printed or handwritten characters from a noisy background is valuable for many applications including test paper autoscoring. The complex structure of Chinese characters makes it difficult to obtain the goal because of easy loss of fine details and overall structure in reconstructed characters. This paper proposes a method for separating Chinese characters based on generative adversarial network (GAN). We used ESRGAN as the basic network structure and applied dilated convolution and a novel loss function that improve the quality of reconstructed characters. Four popular Chinese fonts (Hei, Song, Kai, and Imitation Song) on real data collection were tested, and the proposed design was compared with other semantic segmentation approaches. The experimental results showed that the proposed method effectively separates Chinese characters from noisy background. In particular, our methods achieve better results in terms of Intersection over Union (IoU) and optical character recognition (OCR) accuracy.

Nonstationary Signal Extraction Based on Bat Algorithm Sparse Decomposition Technique

For nonstationary time series i.e. natural electromagnetic field and acoustical signal, effective signal extraction always requires prior knowledge or hypothesis, and hardly do without artificial judgment. We proposed bat algorithm sparse decomposition (BASD) to realize adaptive recognition and extraction of nonstationary signal in a noisy background. We designed two general atomics for typical signals, and developed dictionary training method based on correlation detection and Hilbert transform. The sparse decomposition was turned into an optimizing problem by introducing bat algorithm with optimized fitness function. By contrast with variational modal decomposition, it was indicated that BASD can effectively extract short time target without inducing global aliasing of local feature, and no preset mode number and late screening were needed.

Health Concerns of Textile Workers & Associated Community

Background: Workers in the textile industry risk developing various respiratory and pulmonary diseases due to exposure to cotton dust. The particles from the cotton lint are inhaled by the workers and results in the breathing problems including asthma, shortness of breath, cough and tightness in the chest. The poor health of labor contributes to the low productivity of the labor and in serious cases loss of jobs leading to the poverty. Methods: This study explores the health profiles of the textile workers and associated community and contrast them against the health profile of the control group to factor out any confounding factors. The study is conducted on cotton industry in Kasur, Pakistan. We interviewed 207 workers, 226 people from associated community (living in vicinities of weaving units) and 188 people for control group (from areas far away from weaving units and people are not associated with weaving industry) based on stratified random sampling technique. We employed descriptive methods and logistic regression to explore the association between respiratory diseases and weaving workers. Results: Overall, prevalence of postnasal drip, byssinosis, asthma, and chronic bronchitis were 47%, 35%, 20%, and 10% respectively among the workers. These percentages are significantly higher than the control group. Among workers, 43% & 21% feel difficulty in hearing against noisy background and at low volume respectively. Due to bad light arrangements at workstations, 21% & 31% workers are suffering from myopia and hyperopia respectively. Proportions of the workers suffering from continuous headache, skin infection, depression, and low back pain are 28%, 29%, 27% and 44% respectively. Conclusion: Better environment at workstations, use of protective gears and education are the factors which reduce the risk of associated diseases among workers.

Spatio-temporal and -spectral feature maps in photoplethysmography imaging and infrared thermograph

Background Only a small fraction of the information available is generally used in the majority of camera-based sensing approaches for vital sign monitoring. Dedicated skin pixels, for example, fall into this category while other regions are often disregarded early in the processing chain. Methods We look at a simple processing chain for imaging where a video stream is converted to several other streams to investigate whether other image regions should also be considered. These streams are generated by mapping spatio-temporal and -spectral features of video segments and, thus, compressing the information contained in several seconds of video and encoding these in a new image. Two typical scenarios are provided as examples to study the applicability of these maps: face videos in a laboratory setting and measurements of a baby in the neonatal intensive care unit. Each measurement consists of the synchronous recording of photoplethysmography imaging (PPGI) and infrared thermography (IRT). We report the results of a visual inspection of those maps, evaluate the root mean square (RMS) contrast of foreground and background regions, and use histogram intersections as a tool for similarity measurements. Results The maps allow us to distinguish visually between pulsatile foreground objects and an image background, which is found to be a noisy pattern. Distortions in the maps could be localized and the origin could be discovered. The IRT highlights subject contours for the heart frequency band, while silhouettes show strong signals in PPGI. Reflections and shadows were found to be sources of signals and distortions. We can testify advantages for the use of near-infrared light for PPGI. Furthermore, a difference in RMS contrast for pulsatile and non-pulsatile regions could be demonstrated. Histogram intersections allowed us to differentiate between the background and foreground. Conclusions We introduced new maps for the two sensing modalities and presented an overview for three different wavelength ranges. The maps can be used as a tool for visualizing aspects of the dynamic information hidden in video streams without automation. We propose focusing on an indirect method to detect pulsatile regions by using the noisy background pattern characteristic, for example, based on the histogram approach introduced.