DYNAMIC HANDWRITTEN SIGNATURE IDENTIFICATION USING SPIKING NEURAL NETWORK

The article proposes a method for dynamic signature identification based on a spiking neural network. Three dynamic signature parameters l(t), xy(t), p(t) are used, which are invariant to the signature slope angle, and after their normalization, also to the signature spatial and temporal scales. These dynamic parameters are fed to the spiking neural network for recognition simultaneously in the form of time series without preliminary transformation into a vector of static features, which, on the one hand, simplifies the method due to the absence of complex computational transformation procedures, and on the other hand, prevents the loss of useful information, and therefore increases the accuracy and reliability of signature identification and recognition (especially when recognizing forged signatures that are highly correlated with the genuine). The spiking neural network used has a simple training procedure, and not all neurons of the network are trained, but only the output ones. If it is necessary to add new signatures, it is not necessary to retrain the entire network as a whole, but it is enough to add several output neurons and learn only their connections. In the results of experimental studies of the software implementation of the proposed system, it’s EER = 3.9% was found when identifying skilled forgeries and EER = 0.17% when identifying random forgeries.

Deep semi-supervised learning of dynamics for anomaly detection in laser powder bed fusion

Highly complex data streams from in-situ additive manufacturing (AM) monitoring systems are becoming increasingly prevalent, yet finding physically actionable patterns remains a key challenge. Recent AM literature utilising machine learning methods tend to make predictions about flaws or porosity without considering the dynamical nature of the process. This leads to increases in false detections as useful information about the signal is lost. This study takes a different approach and investigates learning a physical model of the laser powder bed fusion process dynamics. In addition, deep representation learning enables this to be achieved directly from high speed videos. This representation is combined with a predictive state space model which is learned in a semi-supervised manner, requiring only the optimal laser parameter to be characterised. The model, referred to as FlawNet, was exploited to measure offsets between predicted and observed states resulting in a highly robust metric, known as the dynamic signature. This feature also correlated strongly with a global material quality metric, namely porosity. The model achieved state-of-the-art results with a receiver operating characteristic (ROC) area under curve (AUC) of 0.999 when differentiating between optimal and unstable laser parameters. Furthermore, there was a demonstrated potential to detect changes in ultra-dense, 0.1% porosity, materials with an ROC AUC of 0.944, suggesting an ability to detect anomalous events prior to the onset of significant material degradation. The method has merit for the purposes of detecting out of process distributions, while maintaining data efficiency. Subsequently, the generality of the methodology would suggest the solution is applicable to different laser processing systems and can potentially be adapted to a number of different sensing modalities.

Drone-Based Vibration Monitoring and Assessment of Structures

This paper presents a novel method of procuring and processing data for the assessment of civil structures via vibration monitoring. This includes the development of a custom sensor package designed to minimize the size/weight while being fully self-sufficient (i.e., not relying on external power). The developed package is delivered to the structure utilizing a customized Unmanned Aircraft System (UAS), otherwise known as a drone. The sensor package features an electropermanent magnet for securing it to the civil structure while a second magnet is used to secure the package to the drone during flight. The novel B-Spline Impulse Response Function (BIRF) technique was utilized to extract the Dynamic Signature Response (DSR) from the data collected by the sensor package. Experimental results are presented to validate this method and show the feasibility of deploying the sensor package on structures and collecting data valuable for Structural Health Monitoring (SHM) data processing. The advantages and limitations of the proposed techniques are discussed, and recommendations for further developments are made.

An Enhanced Dynamic Signature Verification using the X and Y Histogram Features

Dynamic signature verification by using histogram features is a well-known signature forgery detection technique due to its high performance. However, this technique is often limited to angular histograms derived from vectors containing two adjacent points. We propose additional new features from the X and Y histograms to overcome the limitation. Our experiments indicate that our technique produced Under Curve Area AUC values 0.80 to detect skilled forgery and 0.91 for random forgery. Our method performed best when the verification system uses 12 of the most dominant features. This setup produced AUC values of 0.80 to detect skilled forgery and 0.93 for random forgery. These results outperformed the original technique when the X and Y histogram features are not used that produced AUC values of 0.78 to detect skilled forgery and 0.90 for random forgery.

A Dynamic Transcription Factor Signature Along the Colorectal Adenoma-Carcinoma Sequence in Patients With Co-Occurrent Adenoma and Carcinoma

BackgroundColorectal carcinoma (CRC) often arises from benign adenoma after a stepwise accumulation of genetic alterations. Here, we profiled the dynamic landscapes of transcription factors (TFs) in the mucosa-adenoma-carcinoma progression sequence.MethodsThe transcriptome data of co-occurrent adenoma, carcinoma, and normal mucosa samples were obtained from GSE117606. Identification of differentially expressed TFs (DE-TFs) and subsequent function annotation were conducted in R software. Expression patterns of DE-TFs were clustered by Short Time-series Expression Miner software. Thereafter, modular co-expression analysis, Kaplan-Meier survival analysis, mutation profiling, and gene set enrichment analysis were conducted to investigate TF dynamics in colorectal tumorigenesis. Finally, tissue microarrays, including 51 tumors, 32 adenomas, and 53 normal tissues, were employed to examine the expression of significant candidates by immunohistochemistry staining.ResultsCompared to normal tissues, 20 (in adenoma samples) and 29 (in tumor samples) DE-TFs were identified. During the disease course, 28 expression patterns for DE-TFs and four co-expression modules were clustered. Notably, six DE-TFs, DACH1, GTF2IRD1, MEIS2, NR3C2, SOX9, and SPIB, were identified as having a dynamic signature along the colorectal adenoma-carcinoma sequence. The dynamic signature was of significance in GO enrichment, prognosis, and co-expression analysis. Among the 6-TF signature, the roles of GTF2IRD1, SPIB and NR3C2 in CRC progression are unclear. Immunohistochemistry validation showed that GTF2IRD1 enhanced significantly throughout the mucosa-adenoma-carcinoma sequence, while SPIB and NR3C2 kept decreasing in stroma during the disease course.ConclusionsOur study provided a dynamic 6-TF signature throughout the course of colorectal mucosa-adenoma-carcinoma. These findings deepened the understanding of colorectal cancer pathogenesis.

Kinetochore life histories reveal the origins of chromosome mis-segregation and correction mechanisms

Chromosome mis-segregation during mitosis leads to daughter cells with deviant karyotypes (aneuploidy) and an increased mutational burden through chromothripsis of mis-segregated chromosomes. The rate of mis-segregation and the aneuploidy state are hallmarks of cancer and linked to cancer genome evolution. Errors can manifest as lagging chromosomes in anaphase, although the mechanistic origins and likelihood of correction are incompletely understood. Here we combine lattice light sheet microscopy, endogenous protein labelling and computational analysis to define the life history of >10^4 kinetochores throughout metaphase and anaphase from over 200 cells. By defining the laziness of kinetochores in anaphase, we reveal that chromosomes are at a considerable and continual risk of mis-segregation. We show that the majority of kinetochores are corrected rapidly in early anaphase through an Aurora B dependent process. Moreover, quantitative analyses of the kinetochore life histories reveal a unique dynamic signature of metaphase kinetochore oscillations that forecasts their fate in the subsequent anaphase. We propose that in diploid human cells chromosome segregation is fundamentally error prone, with a new layer of early anaphase error correction required for stable karyotype propagation.