Search

Purpose: Deep learning has made remarkable progress in classifying autism spectrum disorder (ASD) using neuroimaging data. However, the current methods rely mainly on supervised learning, which requires a large amount of manually labeled data, making it an expensive and difficult task to scale. Methods: To overcome this limitation, we propose a novel ensemble-based framework that learns a transferable and generalizable visual representation from different self-supervised features for the downstream task of ASD classification. This framework dynamically learns a superior representation by aggregating complementary information in the frequency domain from independent self-supervised features with limited data. Additionally, to address the information loss caused by the dimensionality reduction of 3D fMRI data, we propose a thresholding algorithm to optimally extract the most discriminant features from 2D rs-fMRI data. Results: Experimental results demonstrate that the proposed method outperforms previous state-of-the-art methods by 19.69% on the ABIDE-1 dataset with a 10-fold cross-validation accuracy of 94.51%. Conclusion: The proposed method learns a transferrable and generalizable ensembled representation by leveraging complementary information encoded in different self-supervised representations for ASD classification.

We present a new open-source and modular instrumentation package composed of up to ten automatic infiltrometers connected to data acquisition systems for automatic recording of multiple infiltration experiments. The infiltrometers are equipped with differential transducers to monitor water level changes in a Mariotte reservoir, and, in turn, to quantify water infiltration rates. The data acquisition systems consist of low-cost components and operate on the open-source microcontroller platform Arduino. The devices were tested both in the laboratory and on different urban and agricultural soils in France and India. More specifically, we tested three procedures to treat the transducers readings, including a filtering algorithm that substantially improved the ability to determine cumulative infiltration from raw data. We combined these three procedures with four methods for estimating the soil parameters from infiltrometer data, showing pros and cons of each scenario. We also demonstrated advantages in using the automatic infiltrometers when infiltration measurements were hindered by: i) linearity in cumulative infiltration curves owing to gravity-driven flow, ii) an imprecise description of the transient state of infiltration, and iii) the occurrence of soil water repellency. The use of the automatic infiltrometers allows the user to obtain more accurate estimates of soil hydraulic parameters, while also reducing the amount of effort needed to run multiple experiments.

We present a new open-source and modular instrumentation package composed of up to ten automatic infiltrometers connected to data acquisition systems for automatic recording of multiple infiltration experiments. The infiltrometers are equipped with differential transducers to monitor water level changes in a Mariotte reservoir, and, in turn, to quantify water infiltration rates. The data acquisition systems consist of low-cost components and operate on the open-source microcontroller platform Arduino. The devices were tested both in the laboratory and on different urban and agricultural soils in France and India. More specifically, we tested three procedures to treat the transducers readings, including a filtering algorithm that substantially improved the ability to determine cumulative infiltration from raw data. We combined these three procedures with four methods for estimating the soil parameters from infiltrometer data, showing pros and cons of each scenario. We also demonstrated advantages in using the automatic infiltrometers when infiltration measurements were hindered by: i) linearity in cumulative infiltration curves owing to gravity-driven flow, ii) an imprecise description of the transient state of infiltration, and iii) the occurrence of soil water repellency. The use of the automatic infiltrometers allows the user to obtain more accurate estimates of soil hydraulic parameters, while also reducing the amount of effort needed to run multiple experiments.

RC4(n, m) is a stream cipher based on RC4 and is designed by G. Gong et al. It can be seen as a generalization of the famous RC4 stream cipher designed by Ron Rivest. The authors of RC4(n, m) claim that the cipher resists all the attacks that are successful against the original RC4. The paper reveals cryptographic weaknesses of the RC4(n, m) stream cipher. We develop two attacks. The first one is based on non-randomness of internal state and allows to distinguish it from a truly random cipher by an algorithm that has access to 24·n bits of the keystream. The second attack exploits low diffusion of bits in the KSA and PRGA algorithms and recovers all bytes of the secret key. This attack works only if the initial value of the cipher can be manipulated. Apart from the secret key, the cipher uses two other inputs, namely, initial value and initial vector. Although these inputs are fixed in the cipher specification, some applications may allow the inputs to be under the attacker control. Assuming that the attacker can control the initial value, we show a distinguisher for the cipher and a secret key recovery attack that for the L-bit secret key, is able to recover it with about (L/n) · 2n steps. The attack has been implemented on a standard PC and can reconstruct the secret key of RC(8, 32) in less than a second.

In this paper, we present three counterfeiting attacks on the block-wise dependent fragile watermarking schemes. We consider vulnerabilities such as the exploitation of a weak correlation among block-wise dependent watermarks to modify valid watermarked %(medical or other digital) images, where they could still be verified as authentic, though they are actually not. Experimental results successfully demonstrate the practicability and consequences of the proposed attacks for some relevant schemes. The development of the proposed attack models can be used as a means to systematically examine the security levels of similar watermarking schemes.

SIN '12 : Fifth International Conference on Security of Information and Networks, 22-27 October 2012, Jaipur, India, Item does not contain fulltext

This work-in-progress paper presents an ensemble-based model for detecting and mitigating Distributed Denial-of-Service (DDoS) attacks, and its partial implementation. The model utilises network traffic analysis and MIB (Management Information Base) server load analysis features for detecting a wide range of network and application layer DDoS attacks and distinguishing them from Flash Events. The proposed model will be evaluated against realistic synthetic network traffic generated using a software-based traffic generator that we have developed as part of this research. In this paper, we summarise our previous work, highlight the current work being undertaken along with preliminary results obtained and outline the future directions of our work.

SIN '12 : Fifth International Conference on Security of Information and Networks, 22-27 October 2012, Jaipur, India, Item does not contain fulltext