Now showing 1 - 4 of 4
  • Publication
    Towards Optimum Counterforensics of Multiple Significant Digits Using Majorisation-Minimisation
    Optimum counterforensics of the first significant digits entails a forger minimally modifying a forgery in such a way that its first significant digits follow some preselected authentic distribution, e.g., Benford’s law. A solution to this problem based on the simplex algorithm was put forward by Comesa Optimum counterforensics of the first significant digits entails a forger minimally modifying a forgery in such a way that its first significant digits follow some preselected authentic distribution, e.g., Benford’s law. A solution to this problem based on the simplex algorithm was put forward by Comesana and Perez-Gonzalez. However their approach requires scaling up the dimensionality of the original problem. As simplex has exponential worst-case complexity, simplex implementations can struggle to cope with medium to large scale problems. These computational issues get compounded by upscaling the problem dimensionality. Furthermore, Benford’s law applies beyond the first significant digit, but no counterforensics method to date offers a solution to handle an arbitrary number of significant digits. As the use of simplex would only aggravate the computational issues in this case, we propose a more scalable approach to counterforensics of multiple significant digits informed by the Majorisation-Minimisation optimisation philosophy.
      473
  • Publication
    Benford's Law: Hammering a Square Peg into a Round Hole?
    Many authors have discussed the reasons why Benford's distribution for the most significant digits is seemingly so widespread. However the discussion is not settled because there is no theorem explaining its prevalence, in particular for naturally occurring scale-invariant data. Here we review Benford's distribution for continuous random variables under scale invariance. The implausibility of strict scale invariance leads us to a generalisation of Benford's distribution based on Pareto variables. This new model is more realistic, because real datasets are more prone to complying with a relaxed, rather than strict, definition of scale invariance. We also argue against forensic detection tests based on the distribution of the most significant digit. To show the arbitrariness of these tests, we give discrete distributions of the first coefficient of a continued fraction which hold in the exact same conditions as Benford's distribution and its generalisation.
    Scopus© Citations 1  157
  • Publication
    Open Source Dataset and Deep Learning Models for Online Digit Gesture Recognition on Touchscreens
    (The Irish Pattern Recognition & Classification Society, 2017-09-01) ; ;
    This paper presents an evaluation of deep neural networks for recognition of digits entered by users on a smartphone touchscreen. A new large dataset of Arabic numerals was collected for training and evaluation of the network. The dataset consists of spatial and temporal touch data recorded for 80 digits entered by 260 users. Two neural network models were investigated. The first model was a 2D convolutional neural (ConvNet) network applied to bitmaps of the glpyhs created by interpolation of the sensed screen touches and its topology is similar to that of previously published models for offline handwriting recognition from scanned images. The second model used a 1D ConvNet architecture but was applied to the sequence of polar vectors connecting the touch points. The models were found to provide accuracies of 98.50% and 95.86%, respectively. The second model was much simpler, providing a reduction in the number of parameters from 1,663,370 to 287,690. The dataset has been made available to the community as an open source resource.
      273
  • Publication
    Automated Identification of Trampoline Skills Using Computer Vision Extracted Pose Estimation
    (Irish Pattern Recognition and Classification Society (IPRCS), 2017-09-01) ; ;
    A novel method to identify trampoline skills using a single video camera is proposed herein. Conventional computer vision techniques are used for identification, estimation, and tracking of the gymnast’s body in a video recording of the routine. For each frame, an open source convolutional neural network is used to estimate the pose of the athlete’s body. Body orientation and joint angle estimates are extracted from these pose estimates. The trajectories of these angle estimates over time are compared with those of labelled reference skills. A nearest neighbour classifier utilising a mean squared error distance metric is used to identify the skill performed. A dataset containing 714 skill examples with 20 distinct skills performed by adult male and female gymnasts was recorded and used for evaluation of the system. The system was found to achieve a skill identification accuracy of 80.7% for the dataset.
      394