IFIP Open Access Digital Library
Not a member yet
22614 research outputs found
Sort by
Simulation Study for Evaluating Efficiency of McPhail Traps in Olive Groves
No full textPart 2: Machine LearningInternational audienceOlive tree production has been of paramount importance to nutrition and culture since the early fifth millennium B.C. The most serious pest of olive groves is the olive fruit fly, known as Bactrocera Oleae or Dacus Oleae, which can lead to loss of production up to 80–90%. Nowadays, measuring the olive fruit fly’s population in the olive groves is key to most pest control strategies. Accordingly, herein, a simulation of olive fruit fly’s population dynamics is presented. Initially, the simulation focuses on the correlation between the ratio of captured olive fruit flies in bait traps in relation to the entirety of population (Trap Efficiency). Subsequently, field and its factors were added in the simulation, such as the ratio of olive fruit flies captured in the traps in relation to flies within the trap’s attraction area (Capture Rate), crops’ variation, and temperature. The simulation’s results initially indicated a correlation between Trap Efficiency and Capture Rate based on previous field experiments, as well as a significant correlation between Trap Efficiency and field Temperature, using various Capture Rates. These results lead towards a contemporary tool for the estimation of olive fruit fly population as well as, by use of regression, the identification of a model that provides trap efficiency estimation for future pests’ traps
Improving Agricultural Image Classification by Mining Images
No full textPart 2: Machine LearningInternational audienceThe task of agricultural image classification has always been a popular topic in agricultural research. Both traditional and deep learning-based methods have emerged to address this task. However, as these methods have expanded, they have become more reliant on data and require additional external information to improve performance. In reality, agricultural images often have low quality and lack annotations, and it is challenging to obtain clear external prior knowledge and semantic information. Therefore, we aim to improve image classification using only the simplest agricultural image dataset, which consists of images and their corresponding class labels. By leveraging the information inherent in the images themselves, we seek to obtain prior knowledge and semantic information to enhance image classification performance, and we use Class Activation Mapping to illustrate the results and the improvement. Furthermore, we enhance the feature extraction process by utilizing it. We conducted experiments on four agriculture-related datasets, using Residual Neural Network as our baseline. The results show that our method achieves improvements in both Top-1 accuracy and Mean Average Precision metrics
Towards Practical Hardware Fingerprinting for Remote Attestation
No full textInternational audienceIn the realm of Trusted Computing for embedded systems, ensuring the integrity of both firmware and hardware presents a complex challenge. Traditional approaches have focused on detecting firmware and operating system (OS) software manipulations, leaving a gap in the identification of subtle hardware modifications and attacks. This paper extends previous work on hardware fingerprinting for remote attestation by conducting and analyzing comprehensive long-term hardware measurements. Building upon the established methodology, we examine the correlation between environmental parameters and analog-to-digital converter (ADC) values to gain suitable reference values for remote attestation procedures. Our work introduces significant contributions: the implementation of two distinct test setups for enhanced hardware fingerprinting, a rigorous evaluation of these measurements to identify strong correlations, the development of a standardized log format for hardware measurements aimed at adoption by the Trusted Computing Group (TCG), and the application to Trusted Platform Module TPM based measured boot and remote attestation. In summary, we integrate hardware manipulation detection with the TPM, and lay the groundwork for a more secure and reliable computing environment in embedded systems
Chain of Trust: Unraveling References Among Common Criteria Certified Products
No full textInternational audienceWith 5394 security certificates of IT products and systems, the Common Criteria for Information Technology Security Evaluation have bred an ecosystem entangled with various kind of relations between the certified products. Yet, the prevalence and nature of dependencies among Common Criteria certified products remains largely unexplored. This study devises a novel method for building the graph of references among the Common Criteria certified products, determining the different contexts of references with a supervised machine-learning algorithm, and measuring how often the references constitute actual dependencies between the certified products. With the help of the resulting reference graph, this work identifies just a dozen of certified components that are relied on by at least 10% of the whole ecosystem – making them a prime target for malicious actors. The impact of their compromise is assessed and potentially problematic references to archived products are discussed
Data Augmentation Techniques for Cross-Domain WiFi CSI-Based Human Activity Recognition
No full textPart 1: Biomedical/ClassificationInternational audienceThe recognition of human activities using WiFi Channel State Information (CSI) facilitates contactless, long-range, and visual privacy-preserving sensing in confined indoor environments. However, the strong environmental dependence inherent to CSI presents a challenge for robust cross-domain generalization, limiting its practical applicability. Drastic environmental variations, such as transitions between line-of-sight (LOS) and non-line-of-sight (NLOS) through-wall scenarios or changes in antenna configurations, introduce a significant domain gap that can lead to severely degraded model performance at test time. To address the challenge of model generalization in these demanding cross-scenario and cross-system settings, an area that remains under-explored, this work investigates the effectiveness of data augmentation techniques commonly utilized in image-based learning when applied to WiFi CSI. We collect and make publicly available a dataset of CSI amplitude spectrograms of human activities. Utilizing this data, an ablation study is conducted in which we train activity recognition models based on the EfficientNetV2 architecture, allowing us to evaluate the impact of each augmentation on model generalization performance. The results show that, although no single technique is universally effective, specific combinations of data augmentations applied to CSI amplitude features can significantly enhance generalization in certain cross-scenario and cross-system settings
Vision Transformer Based Tokenization for Enhanced Breast Cancer Histopathological Images Classification
No full textPart 1: Biomedical/ClassificationInternational audienceBreast cancer remains a global concern, underscoring the crucial need for early diagnosis to ensure effective treatment. In recent years, convolutional neural networks (CNNs) have dominated medical vision tasks, but recent interest in computer-aided diagnosis (CAD) has shifted towards vision transformers (ViTs). However, the vision transformer (ViT) is recognized for its data-intensive nature and a substantial number of parameters, resulting in poorer performance compared to CNNs. The challenges associated with the data-intensive characteristics and extensive parameters of ViTs are especially pronounced in tasks involving medical image datasets characterized by data scarcity. To address this gap, our paper proposes the TokenLearner model for classifying breast tumours in histopathological images. This hybrid model fuses ViT and convolution layers operating directly on input patches and utilizing standard convolutions in the attention map block to dynamically highlight relevant regions in the input patches, minimizing the number of patches used in training for a Vision Transformer with lower training time and complexity compared to the ViT base architecture. The model was extensively evaluated using the BreakHis dataset, comprising 2496 benign and 5429 malignant masses, resulting in remarkable performance. It achieved an accuracy of 97.04%, precision of 96.99%, sensitivity of 96.11%, and F1 score of 96.54% for breast mass classification. Our study highlights the effective utilization of trained attention mechanisms in developing high- performance computer-aided systems for breast cancer diagnosis. Importantly, this approach achieves outstanding results while minimizing computational resource requirements and reducing processing time
Modeling Distributed and Flexible PHM Framework Based on the Belief Function Theory
No full textPart 1: Biomedical/ClassificationInternational audienceThis paper explores the integration of the belief function theory within the domain of Prognostics and Health Management (PHM), offering a novel approach to decision-making under conditions of uncertainty and incomplete information. Central to our methodology is the modeling of beliefs and uncertainties through belief mass functions, enabling the representation and aggregation of diverse information source. This approach is particularly advantageous in dynamic and complex environments characteristic of PHM, where data may be partial or conflicting. By adjusting the weight of information based on source reliability, our framework supports nuanced and adaptive decision-making. Furthermore, the proposed model facilitates collaborative decision-making in distributed systems by effectively managing information diversity and resolving conflicts. While focused on PHM, the versatility of our approach, experimented through numerical examples, allows for potential applications across various fields requiring robust and adaptive decision-making strategies in the face of uncertainty
A privacy-preserving graph encryption scheme based on oblivious RAM
No full textPart 2: Crypto ApplicationInternational audienceGraph encryption schemes play a crucial role in facilitating secure queries on encrypted graphs hosted on untrusted servers. With applications spanning navigation systems, network topology, and social networks, the need to safeguard sensitive data becomes paramount. Existing graph encryption methods, however, exhibit vulnerabilities by inadvertently revealing aspects of the graph structure and query patterns, posing threats to security and privacy. In response, we propose a novel graph encryption scheme designed to mitigate access pattern and query pattern leakage through the integration of oblivious RAM and trusted execution environment techniques, exemplified by a Trusted Execution Environment (TEE). Our solution establishes two key security objectives: (1) ensuring that adversaries, when presented with an encrypted graph, remain oblivious to any information regarding the underlying graph, and (2) achieving query indistinguishability by concealing access patterns. Additionally, we conducted experimentation to evaluate the efficiency of the proposed schemes when dealing with real-world location navigation services
Guiding Process Mining Projects with the IPMM Framework: A Case Study with a German Manufacturer
No full textPart 1: Digital Transformation Approaches in Production and ManagementInternational audienceProcess mining is a transformative approach that leverages event data from disparate information systems to discover, improve performance, and ensure compliance of business processes. Bridging the gap between theory and practice, especially for newcomers, is a pressing concern. To address this issue, we have developed our Integrative Process Mining Management (IPMM) framework, which consists of a methodology and tools to provide organizations with a concise overview of existing process mining methodologies, their pros and cons, and their alignment with current maturity levels. Within our methodology we present our management tool: The Process Mining Maturity Level Cheat Sheet. This process mining management tool outlines potential goals for each maturity level. The IPMM framework offers a structured solution. Our framework emphasizes the importance of tailoring process mining techniques to an organization’s maturity level. Finally, we apply our framework in a German manufacturing company
Information Security Theory and Practice: 14th IFIP WG 11.2 International Conference, WISTP 2024, Paris, France, February 29 – March 1, 2024, Proceedings
No full textInternational audienceThis volume constitutes the refereed proceedings of the 14th IFIP WG 11.2 International Conference on Information Security Theory and Practices, WISTP 2024, held in Paris, France.The 12 full papers presented were carefully reviewed and selected from 30 submissions. The papers presented in this proceedings focus on emerging trends in security and privacy, including experimental studies of fielded systems while exploring the application of security technology, and highlighting successful system implementations