ICATA 24 Conference

On the use of Transfer Learning to Improve Breast Cancer Detection

Khadija Aguerchi , Younes jabrane, Maryam habba ,Mustapha Ameur

ABSTRACT

Breast cancer is one of the most common malignancies in women globally, and early identification is critical for better patient outcomes. Deep learning has developed in recent years as a promising approach for automating the identification of breast cancer in mammograms. Transfer learning, which involves adapting a pre-trained model to a new task, is a promising method for enhancing the efficiency and accuracy of breast cancer diagnosis using deep learning. This work studies the efficacy of transfer learning strategies in detecting breast cancer using pre-trained deep-learning models. Using large mammographic datasets, we investigate several transfer learning algorithms and assess their effects on detection performance metrics such as accuracy, precision, recall and ROC AUC. The study's results enhance automated breast cancer detection and shed light on how well transfer learning strategies can improve the precision and dependability of detection.

Keywords: Breast cancer, transfer learning, medical imaging, deep learning.

DOI: https://doi.org/10.5281/zenodo.11217339

Efficient Service Composition Optimization under Uncertain Environment

REMACI Zeyneb Yasmina

Abstract

Over the past few years, companies have been offring a variety of functionalities through Web services as part to stay competitive. Meanwhile, the customers may face challenges in choosing a service that meets their requirements due to the similarity of functionalities. Moreover, the customers’ requests often emerge with a complex nature in an uncertain environment that is rarely fulfilled by an atomic service. Hence, there is a necessity for optimizing service composition based on Quality of Service (QoS) with taking into account environmental uncertainty. To attain the mentioned goal, our proposed approach adopts a local and global optimization strategies. Firstly, we adopt a heuristic based on the Entropy, Cross-Entropy, and the deviation degree for the hesitant fuzzy set to rank similar Web services. Afterwards, we introduce an improved metaheuristic called Group Leaning based Composition as a global optimization for selecting the near-optimal composition.

Key words: Deviation degree, Global Optimization, Group learning metaheuristic, Hesitant fuzzy set, Local Optimization, Service Composition, Quality of Service, Web Service.

DOI:https://doi.org/10.5281/zenodo.11217801

Understanding User Intention in Pervasive Environments: A Literature Review and Perspectives

Abdelhadi Bouain, Mohamed Nezar ABOURRAJA, Mohamed Yassine Samiri

Abstract

Identifying user intention (what the user wishes to achieve within a system) with minimal or ideally no direct user interaction is a major goal in pervasive computing. Achieving this goal requires a clear and consistent definition of intention, a concept widely used but understood differently across various studies. In this work, we first aim to clarify the different interpretations of intention, distinguishing between implicit and explicit intention. Subsequently, we compare various existing approaches from the literature, seeking to reconcile these diverse viewpoints and establish a common foundation for future research efforts.

Key words: Pervasive computing, User intention, User intent prediction, Multimodal Large Language Models.

DOI: https://doi.org/10.5281/zenodo.11237705

Hand Gesture Recognition Using Convolutional Neural Networks

Veluru Karthik Reddy, Vanapalli Durga Prasanth, R.Shiva rama Krishna, Naidu Sri lekha, and Jyothi N.M

Abstract

Hand gestures play a crucial role in communication and are essential in various scenarios where verbal communication is not possible. For instance, traffic policemen, newsreaders, airport staff, and gamers often rely on hand signals to communicate. Therefore, there is a growing need for robust hand pose recognition (HPR) methods that can identify hand gestures accurately. However, the current state-of-the- art HPR methods struggle with identifying hand gestures in the presence of cluttered backgrounds .To address this challenge, we propose a deep learning framework based on convolutional neural networks (CNNs) to identify hand postures regardless of hand size, location in the image, and background clutter. Our proposed CNN-based approach eliminates the need for feature extraction and learns to recognize hand poses without explicit foreground segmentation. This method effectively identifies hand poses, even in the presence of complex and varying backgrounds or poor lighting conditions .We have conducted several experiments, which demonstrate the superiority of our proposed method over state-of-the-art datasets. Our approach significantly improves the accuracy of hand pose recognition, making it more reliable and efficient for a wide range of applications. Therefore, our proposed method has significant potential for use in real-world scenarios, such as traffic management, sign language interpretation, and virtual reality gaming .Overall, our results suggest that deep neural networks can provide a robust and effective solution for hand gesture recognition tasks.

Key words: HPR , CNN, Segmentation ,Background clutter, Virtual Reality ,Neural Networks.

DOI: https://doi.org/10.5281/zenodo.11237998

Eating Smart: Advancing Health Informatics with the Grounding DINO-based Dietary Assistant App 

Abdelilah Nossair, Hamza El Housni

Abstract

The Smart Dietary Assistant project combines technology and Machine Learning (ML) to offer personalized advice for people with dietary concerns such as diabetes. This approach focuses on the user helping them make decisions about their diet using the Grounding DINO model. Grounding DINO uses a text encoder and image backbone to improve detection accuracy without relying on a labeled dataset making it practical for real world situations with various food types. This model uses a 52.5 AP score on the COCO dataset and attention mechanisms that leverage features based on user-provided labels and food images to allow precise object recognition. The feature is at the core of the user app, turning smartphones into a helpful dietary advisor that enables people to manage their health effectively.

The app can use your device camera to take photos that will be analyzed by the model for detection and categorize the food items correctly. This is what differs in this system: it decides to be free and not to be connected to annoying cloud databases of information. The application uses a database managed by itself that is of PostgreSQL type, ensuring the preservation of data integrity and control. This database hosting information includes all types of food products, from profiles to health insights drawn from their consumption by human beings. This helps in effective and efficient data access speed, reliability, and enhances user privacy through localized storage within the organizational infrastructure.

The app focuses on improving the experiences of the users, considering that it allows them to create profiles through which they describe themselves based on preferences and tips on nutrition. In addition to calories information, the app provides insights to nutrients such as proteins, vitamins, and minerals. This makes it possible for one to decide the kind of food to take, either for weight management, muscle building, or managing health conditions. On the other part, it also assesses food compatibility versus profiles and gives personal recommendations for alternatives and recipes. Such kind of personal help is highly convenient for persons with needs as it helps them take their healthy options confidently.

Developed using React Native and TypeScript, the Smart Dietary Assistant app guarantees operation across devices and platforms. It incorporates technologies beyond modeling to ensure optimal performance in food recognition, scalability for future enhancements and seamless integration, with other dietary tools.

Users have the option to enjoy features like using the camera to scan food items, for tracking habits and receiving insightful analysis. They can also interact with an assistant for recommendations. The protection of data is ensured through user authentication whereas customizable settings enhance the user experience. React Native enables smooth screen transitions. The expo camera allows scanning capabilities. Local storage efficiently manages data to create an easy/appealing to use interface.

The Smart Dietary Assistant app’s interface stands out for striking a balance between aesthetics and usability. The use of buttons, and a vibrant color scheme enhances user experience by making navigation and feature selection simple. The chatbot feature, represented by an avatar encourages user engagement and personalized guidance seeking. Users find camera scanning convenient although it is noted that varying lighting conditions may affect accuracy. It is this appreciation that opened doors to improvement that can guarantee success in all situations.

The choice of a self-hosted PostgreSQL database for this project re-emphasizes its importance in the realms of health informatics and nutritional science. This is data that can be stored without really depending on outside cloud services, and just with that, the same can be retained as reliable information, since there are chances that it can be changed from the outside.

In the future, the Smart Dietary Assistant is planned to be empowered with collaboration with devices. With this development, the application can sync with fitness trackers and smartwatches to give time-based suggestions from physiological data such as blood sugar level and calories burnt. This will connect users to devices that give them individualized advice regarding their health needs, depending on the style of activity. The application is open to collaborations with AI-powered tools in the development of personalized recipes and meal plans that would give the user an easy time adhering to his preferences, dietary restrictions, and time-in sync physiological information. With conditions like diabetes, this holistic approach to diet management is deemed beneficial because it would make the app utilities more effective, always supports objectives for weight management or muscle building, and therefore supports the overall well-being of the user.

Key words: Food Image Recognition, Machine Learning in Nutrition, Zero-Shot Object Detection

DOI: https://doi.org/10.5281/zenodo.11243881

A decision support system driven by artificial intelligence for industrial applications

Eng. Hala Mellouli, Prof. Anwar Meddaoui, Prof. Abdelhamid Zaki

Abstract

Decision-making in industrial settings is a continuous process that drives the organization's overall performance. It implies consistently selecting the optimal alternative, regularly reviewing the effectiveness of the decision, learning from its consequences, and refining the decision-making framework accordingly. in the modern era, characterized by the abundance of data, the ineffectiveness of conventional multi-criteria decision-making methods to process large volumes of data prevails over their ability to manage the multidimensional nature of decision-making in industrial settings, hence to cope with the increasing complexity of process industrials are challenged to explore the potential of artificial intelligence to optimize their decisions. In the current work, a new decision-making approach is introduced, the model combines artificial neural networks with the Analytic Hierarchy Process and the balanced scorecard to provide real-time decision-making recommendations for complex industrial problems.

Key words: Industrial performance, Artificial Neural Network, Analytical Hierarchy Process, Decision support system.

DOI:https://doi.org/10.5281/zenodo.11244054

Semantic Enrichment of Moroccan Tourism Ontology Based NLP Technologies

ASMA AMALKI, KHALID TATANE, ALI BOUZIT

Abstract

Ontology is a conceptual representation model that allows sharing and reuse of a domain knowledge in a human and machine-readable format. However, the massive amount of knowledge available today makes ontology enrichment a challenging task. In this paper, we present a semi-automatic approach of ontology learning from collection of domain specific texts, to text processing based NLP tools, to enrich an existing ontology. This study utilize data crawling from official websites. Concepts and relations extraction is done automatically from a textual corpus and domain experts do consistency and redundancy check manually. The proposed approach is applied in order to enrich Moroccan tourism ontology named OTM with semantic entities extracted from heterogeneous data sources.

Key words: Corpus, Knowledge acquisition, NLP, ontology enrichment, ontology learning.

DOI:https://doi.org/10.5281/zenodo.11244423

New Compression Point Reducing Memory Size in Field of Characteristic Different From 2 And 3

ASSOUJAA ISMAIL, EZZOUAK SIHAM

Abstract

Compression point is a new method to compress the space memory and still have the same data. In this paper, we will present a new method of compression points work well with addition opera- tion in elliptic curve, so instead of storing the value of two points P = (xP , yP ), Q = (xQ , yQ ), we will store the addition of the x- coordinates i,e (α = xP + xQ , yP , yQ ) or the y-coordinates i,e (xP , xQ , β = yP + yQ). In this article, we show a new technique for compressing two points in elliptic curve with different coordinate system: Affine, Projective and Jacobian in a field of characteristic ≠ 2 & 3, and show the cost of theses operations. This method can save if we work with affine, Projective or Jacobian coordinates, at least 25%, 17%, 17% of memory size respectively, and also see what happens in case if we take Edwards curve and Montgomery curve cases.

Key words: Elliptic curve, Affine coordinate, Projective coordinate, Jacobian coordinate, compression point.

DOI: https://doi.org/10.5281/zenodo.11244720

Digitizing Sports Arenas: New Trends and Technologies

JIBRAILI Malak

Abstract

The digital revolution has transformed the world we live in, including the professional realm, which has harnessed its advantages. Digital technologies are widespread in the sports world, used daily in all its facets: strategically employed by brands, leveraged by coaches and athletes to optimize performance. It's worth noting that the sports spectacle accompanies and even inspires technical innovations (Wille, 2015)1, and the adoption of these tools provides a decisive "competitive advantage," both in sports and commercially (Giblin et al., 2016)2. For the field of sports marketing, the rise of digital offers fervent and loyal supporters, the famous fans, a new way to experience current sports events and immerse themselves in them (Bailey, 2017)3. Today, we talk about a "connected match," a "smart stadium," and "digital sports." It is essential to understand how the introduction of technological tools within sports facilities, especially in the behavior of fans, positively or negatively changes the landscape. A literature review outlines the technological tools introduced in sports venues, the role of social media as a new communication channel, and the digital's position in the sports entertainment industry. A qualitative study then confronts the views of sports managers with those of fans regarding the role of various technological tools within sports facilities.

Key words: Sports Venue, Sports Industry, Technologies, Digital

DOI: https://doi.org/10.5281/zenodo.11245185

Intelligent security system for access doors based on the implementation of CNN on FPGA board

Bouchra KOUACH, Mohcin Mekhfioui, Rachid El gouri

Abstract

Nowadays, there's a great deal of interest in the security of digital data, but it's crucial to recognize that the physical protection of equipment containing sensitive information is essential. Indeed, the loss of a server is the same as the loss of data. With this in mind, this article proposes an intelligent security system for datacenter room access doors, to guarantee optimum protection. To this end, this work seeks to implement a door security system based on image classification by the CNN convolutional neural network applied to the FPGA board. Our approach is to use deep learning frameworks such as TensorFlow and Keras for model development and training. In terms of FPGA type, we're using the ZYNQ board. To deploy this accelerated CNN model, we use PetaLinux. The results of this work demonstrate that there is an energy saving over the traditional CPU or GPU implementation of CNN. In addition, FPGAs are well adapted to the parallel nature of CNN computations in real-time applications.

Key words: CNN, FPGA, Zynq, PNQY, PetaLinux, Tensorflow, Keras, Image classification, Python.

DOI: https://doi.org/10.5281/zenodo.11245434

Copyright Law Governing Works Resulting from the Use of Generative AI and Emerging Jurisprudence in USA and China.

M.P.Ramaswamy 

Abstract

Legal and regulatory challenges facing generative Artificial Intelligence (AI) gain significance due to the indispensable need for legal certainty, which is essential for developing and deploying generative AI. Although significant economies worldwide have formulated aspiring policies for leadership in AI development, specific initiatives to create a congenial legal environment are sparse. Despite some regional initiatives and national regulatory standards are starting to emerge, lack of harmonization and a concerted approach are conspicuous. While more international initiatives to establish legal standards governing AI are expected, many national jurisdictions still rely on their existing general legal regimes to govern AI-related developments. However, with the faster growth of generative AI and its wider adoption and accessibility across borders, national Courts are increasingly confronted with unprecedented legal claims and regulatory issues. The paper assesses some legislative measures and early judicial responses pertinent to generative AI. It identifies evidence of some drastic diversity in national responses and argues its potential implications impeding the development and use of generative AI in key economies. The paper specifically focuses on the intellectual property rights (IPR) issues related to copyrights in works produced by generative AI. It compares the two significant jurisdictions of China and the USA as representations of civil and common law legal systems, respectively. The paper reviews the copyright regimes and legal standards pertinent to the use of generative AI in the two jurisdictions. It closely examines the emerging judicial jurisprudence in the two jurisdictions tackling copyright claims related to generative AI and identifies its potential implications. The paper concludes with an assessment of the comparative legal findings in the two jurisdictions to determine whether there are any substantive differences that emerge and recommends certain measures for a cohesive approach in the addressing IPR challenges facing generative AI.

Key words: Copyrights, AI Works, National Legal Standards, Judicial Approach, Comparative Law

DOI: https://doi.org/10.5281/zenodo.11526273

Lean Management in Correlation with the IATF 16949:2016 Automotive Standard: What Impact on Automotive Companies’ Performance?

Oumaima EL AFFAKI, Mariam BENHADOU, Abdellah HADDOUT

Abstract

Facing the competitiveness and evolution of the sector, automotive manufacturing companies are concentrating their efforts on producing components at the required level of quality and the best cost while respecting delivery times. Improving systems and performance, optimizing processes, controlling risks, seizing opportunities, and achieving targets are also among their priorities.  The “Lean” approach derives from the Toyota production system; it aims to maximize the gains of organizations while minimizing costs. Lean Management is a set of principles, tools, and methods that eliminate waste throughout the supply chain, optimize the operation of different processes, and improve effectiveness, efficiency, and business performance while involving personnel and creating a teamwork spirit. Lean Manufacturing is adopted by operational processes within manufacturing companies, it aims to eliminate non-value-added operations related to overproduction, inventory, waiting time, motion, transportation, defects and errors, overprocessing and non-utilized talent. Nowadays, original equipment manufacturers as well as automotive companies of different tiers require their suppliers to deliver just-in-time (JIT), to manufacture products conforming to predefined characteristics, to reduce costs, and to achieve defined quality objectives. In order to ensure their sustainability and with the aim of continually meeting customer needs, automotive companies are adopting Lean Management tools and methods. The objective of Lean Management is to optimize flows, improve operational processes’ efficiency, satisfy customers, ensure the conformity of products and services, deliver on time, and minimize related costs. This is in alliance with clause 10.3.1 of the international automotive standard, which requires organizations to implement a continuous improvement approach in order to reduce waste and optimize the functioning of the automotive supply chain. On the other hand, IATF 16949:2016 is the international standard that defines the quality management systems (QMS) requirements for design and development, manufacturing, assembly, installation, and service intended for organizations operating in the automotive sector. The standard is applicable in the automotive supply chain regardless of the supplier’s tier and the nature of the supplied product.  IATF 16949:2016 follows the "HLS: High Level Structure", and its requirements are organized around 10 chapters structured according to the PDCA (Plan-Do-Check-Act) cycle.  As per the ISO 9001:2015 standard, the first three chapters of the IATF define the scope of the standard and the vocabulary specific to the automotive industry, and the following seven contain the detailed requirements. “Plan” step encompasses Chapter 4: “Context of the organization”, Chapter 5: “Leadership”, Chapter 6: “Planning”, and Chapter 7: “Support”.  The “Do” step contains operational requirements dictated in Chapter 8: “operation”. Chapter 9: “Performance evaluation” is assigned to the “check” step, and requirements related to Chapter 10: “improvement” are allocated to the “act” stage.       The certification of QMS according to the international automotive standard IATF 16949:2016 is required by the automotive market, customers, and automotive manufacturers. That is why the QMS certification is one of automotive suppliers’ top priorities; it is considered a license to operate in the automotive market. To ensure the organizations’ sustainability, automotive suppliers should certify their QMS to the IATF standard while ensuring the integration of its requirements into their system and maintaining compliance with IATF clauses continuously. Some studies have demonstrated that Lean Management excels in the automotive industry due to high demand and competition, as well as specific customer requirements and the requirements dictated in the international automotive standard IATF 16949:2016. The latter requires organizations to define a continuous improvement process and to deploy particular efforts to reduce waste and increase customer satisfaction. Through its requirements, IATF 16949:2016 encourages automotive organizations to develop a QMS capable of providing continuous improvement, the prevention of quality defects, the reduction of waste and variations throughout the supply chain, and continuing to meet the demands of their customers. In this context, the present paper provides guidance to automotive suppliers in order to improve their performance results by analyzing and defining the impact of the adoption of Lean Management tools and the compliance of QMS with IATF 16949:2016 requirements on performance improvement. For this purpose, a correlation analysis was performed between the Lean Management concept and the IATF 16949:2016 requirements. The contribution of each Lean Management tool and the related IATF requirement to operational performance improvement is defined in the present study. The considered key performance indicators (KPIs) were determined according to the IATF standard, and we classify them into two categories: internal KPIs and external KPIs.  The cost of internal poor quality, process effectiveness, process efficiency, product conformance, and maintenance performance are part of the internal KPIs, while delivered part quality performance, delivery performance, customer disruptions, warranty, and the cost of external poor quality evaluate the external performance. The findings show that a substantial improvement in both internal and external performance as well as customer satisfaction can be achieved through the implementation of Lean Management tools in conjunction with the QMS conformance with the international automotive standard requirements.

Key words: automotive industry, IATF 16949, key performance indicator, Lean management, quality management system.

DOI: https://doi.org/10.5281/zenodo.11245690