Clients are increasingly looking for fast and effective means to quickly and frequently survey and communicate the condition of their buildings so that essential repairs and maintenance work can be done in a proactive and timely manner before it becomes too dangerous and expensive. Traditional methods for this type of work commonly comprise of engaging building surveyors to undertake a condition assessment which involves a lengthy site inspection to produce a systematic recording of the physical condition of the building elements, including cost estimates of immediate and projected long-term costs of renewal, repair and maintenance of the building. Current asset condition assessment procedures are extensively time consuming, laborious, and expensive and pose health and safety threats to surveyors, particularly at height and roof levels which are difficult to access. This paper aims at evaluating the application of convolutional neural networks (CNN) towards an automated detection and localisation of key building defects, e.g., mould, deterioration, and stain, from images. The proposed model is based on pre-trained CNN classifier of VGG-16 (later compaired with ResNet-50, and Inception models), with class activation mapping (CAM) for object localisation. The challenges and limitations of the model in real-life applications have been identified. The proposed model has proven to be robust and able to accurately detect and localise building defects. The approach is being developed with the potential to scale-up and further advance to support automated detection of defects and deterioration of buildings in real-time using mobile devices and drones.

In the field of supervised machine learning, the quality of a classifier model is directly correlated with the quality of the data that is used to train the model. The presence of unwanted outliers in the data could significantly reduce the accuracy of a model or, even worse, result in a biased model leading to an inaccurate classification. Identifying the presence of outliers and eliminating them is, therefore, crucial for building good quality training datasets. Pre-processing procedures for dealing with missing and outlier data, commonly known as feature engineering, are standard practice in machine learning problems. They help to make better assumptions about the data and also prepare datasets in a way that best expose the underlying problem to the machine learning algorithms. In this work, we propose a multistage method for detecting and removing outliers in high-dimensional data. Our proposed method is based on utilising a technique called t-distributed stochastic neighbour embedding (t-SNE) to reduce high-dimensional map of features into a lower, two-dimensional, probability density distribution and then use a simple descriptive statistical method called interquartile range (IQR) to identifying any outlier values from the density distribution of the features. t-SNE is a machine learning algorithm and a nonlinear dimensionality reduction technique well-suited for embedding high-dimensional data for visualisation in a low-dimensional space of two or three dimensions. We applied this method on a dataset containing images for training a convolutional neural network model (ConvNet) for an image classification problem. The dataset contains four different classes of images: three classes contain defects in construction (mould, stain, and paint deterioration) and a no-defect class (normal). We used the transfer learning technique to modify a pre-trained VGG-16 model. We used this model as a feature extractor and as a benchmark to evaluate our method. We have shown that, when using this method, we can identify and remove the outlier images in the dataset. After removing the outlier images from the dataset and re-training the VGG-16 model, the results have also shown that the accuracy of the classification has significantly improved and the number of misclassified cases has also dropped. While many feature engineering techniques for handling missing and outlier data are common in predictive machine learning problems involving numerical or categorical data, there is little work on developing techniques for handling outliers in high-dimensional data which can be used to improve the quality of machine learning problems involving images such as ConvNet models for image classification and object detection problems.

Clients are increasingly looking for fast and effective means to quickly and frequently survey and communicate the condition of their buildings so that essential repairs and maintenance work can be done in a proactive and timely manner before it becomes too dangerous and expensive. Traditional methods for this type of work commonly comprise of engaging building surveyors to undertake a condition assessment which involves a lengthy site inspection to produce a systematic recording of the physical condition of the building elements, including cost estimates of immediate and projected long-term costs of renewal, repair and maintenance of the building. Current asset condition assessment procedures are extensively time consuming, laborious, and expensive and pose health and safety threats to surveyors, particularly at height and roof levels which are difficult to access. This paper aims at evaluating the application of convolutional neural networks (CNN) towards an automated detection and localisation of key building defects, e.g., mould, deterioration, and stain, from images. The proposed model is based on pre-trained CNN classifier of VGG-16 (later compaired with ResNet-50, and Inception models), with class activation mapping (CAM) for object localisation. The challenges and limitations of the model in real-life applications have been identified. The proposed model has proven to be robust and able to accurately detect and localise building defects. The approach is being developed with the potential to scale-up and further advance to support automated detection of defects and deterioration of buildings in real-time using mobile devices and drones.

Condition assessment and health monitoring (CAHM) of built assets requires effective and continuous monitoring of any changes to the material and/or geometric properties of the assets in order to detect any early signs of defects or damage and act on time. Most of the traditional CAHM techniques, however, depend on manual labour despite that, in some cases, the inspection environment can be unsafe and could lead to low efficiency or misjudgement of the severity of the defect. In recent years, computer vision techniques have been proposed as an automated alternative to the traditional CAHM techniques as methods for extracting and analysing feature-related information from asset images and videos. Such methods have proven to be robust and effective solutions, complementary to current time-consuming and unreliable manual observational practices. This work is concerned with the development of a deep learning-based smartphone app, which allows real-time detection of four types of defects in buildings, namely, cracks, mould, stain and paint deterioration. Since smartphones are widely available and equipped with highresolution cameras, this application can offer a practical, low-cost solution for condition assessment procedures of built assets. The obtained results are promising and support the feasibility and effectiveness of the approach to identify and classify various types of building defects. K E Y W O R D Scondition assessment and health monitoring, deep learning, defects in buildings, object detection, smartphone applications | INTRODUCTIONDefects are defined as the deterioration of building features and services to unsatisfactory quality levels of the requirements of the users. 1 Defects and deterioration are common problems in built structures; they are a major concern of buildings and require significant attention. Although various defects are more common in old buildings, they can occur in new buildings as well. 2 Workers' faults, natural causes, poor maintenance and the age of the building are amongst the common causes of defects in built structures. 3 There are many types of building defects. In general, these can be divided into two categories: structural and non-structural defects. Structural defects are those which occur in a

We present a three-dimensional non-iterative reconstruction algorithm developed for conductivity imaging with real data collected on a planar rectangular array of electrodes. Such an electrode configuration as well as the proposed imaging technique are intended to be used for breast cancer detection. The algorithm is based on linearizing the conductivity about a constant value and allows real-time reconstructions. The performance of the algorithm was tested on numerically simulated data and we successfully detected small inclusions with conductivities three or four times the background lying beneath the data collection surface. The results were fairly stable with respect to the noise level in the data and displayed very good spatial resolution in the plane of electrodes.

BACKGROUND: Genome-wide association studies (GWASs) have enriched the fields of genomics and drug development. Adrenocortical carcinoma (ACC) is a rare cancer with a bimodal age distribution and inadequate treatment options. Paediatric ACC is frequently associated with TP53 mutations, with particularly high incidence in Southern Brazil due to the TP53 p.R337H (R337H) germline mutation. The heterogeneous risk among carriers suggests other genetic modifiers could exist. METHODS: We analysed clinical, genotype and gene expression data derived from paediatric ACC, R337H carriers, and adult ACC patients. We restricted our analyses to single nucleotide polymorphisms (SNPs) previously identified in GWASs to associate with disease or human traits. RESULTS: A SNP, rs971074, in the alcohol dehydrogenase 7 gene significantly and reproducibly associated with allelic differences in ACC age-of-onset in both cohorts. Patients homozygous for the minor allele were diagnosed up to 16 years earlier. This SNP resides in a gene involved in the retinoic acid (RA) pathway and patients with differing levels of RA pathway gene expression in their tumours associate with differential ACC progression. CONCLUSIONS: These results identify a novel genetic component to ACC development that resides in the retinoic acid pathway, thereby informing strategies to develop management, preventive and therapeutic treatments for ACC.

Based on The National Electrical Manufacturers Association (NEMA), using the AMINE software to construction of sinograms and using a positron emission source of 22 Na, were made calculations to determine the spatial resolution of a ring array system of phoswich detectors of positron emission tomograph included in the CLEAR PET-XPAD3/CT prototype for small animals, made in the laboratories of CCPM and whose project is led by the research group ImXgam. The radioactive source 22 Na of approximately 9 MBq of activity, with spherical shape and diameter of 0.57mm is immersed in a plexiglas disc that was located at the geometric center of tomographic system with a Field of View (FOV) of 35 mm in the axial and transverse directions. Displacements of radioactive source were performed on the three cartesian axes and was rebuilt a sinogram for each axis. The shape of sinogram allows describe the correct position and the maximum efficiency of each detector. Subsequently, was carried out a scanning in each one of three spatial axes taking enough distance to cover the dimensions of radioactive source. Data for each phoswich detector were recorded. The process was repeated for other axes and then radioactive source was centered with respect to the FOV and were calculated FWHM (Full Width at Half Maximum) and FWTM (Full Width at Tenth Maximum) values and performing statistics of these values with parabolic fitting, the latter setting allows to obtain parameters of spatial resolution of system.

The structural health monitoring (SHM) of buildings is a crucial task which requires continuous monitoring of the physical properties of the asset to obtain actionable and quantitative information about the structure over time. Traditional SHM techniques, however, are tedious and depend on manual labour. Moreover, the inspection of unsafe environment can be ineffective and may lead to miss-judgement of the severity of the damage. The use of Unmanned Aerial Vehicles (UAVs), in recent years, as an alternative survey method has fundamentally revolutionised the way health monitoring and damage assessment for engineering structures are carried out. UAVs, otherwise known as drones, do not require the presence of people around the areas where inspection is taking place. A drone can also be remotely maneuvered very effectively and can be taken to regions that are characterized by difficult accessibility or pose dangerous risks. Owing to these unique capabilities, the automatic drone-based monitoring of structural assets has proven to offer efficient, reliable, and high integrity inspection even in the most challenging diverse environments. This work is devoted to the development of a deep learning-based drone system suitable for real-time detection of structure damages. The drone system is equipped with an on-broad single board computer and a high definition camera. The on-board computer transmits real-time videos captured by the drone camera over WIFI connection to a GPU-based ground station running a deep-learning neural network for damages detection. This application can offer a practical, effective and low-cost solution for condition assessment procedures of structural systems. The obtained results support the feasibility and effectiveness of the approach to identify and classify various types of damages in real-time, using the detection of rust on metal roofs as a problem domain.