Fouad Bousetouane, Ph.D

Fouad Bousetouane, Ph.D

Greater Chicago Area
7K followers 500+ connections

Join to view profile

About

I am an AI expert and technical leader with a focus on generative AI and multimodal…

Contributions

Join now to see all contributions

Activity

Join now to see all activity

Experience

  • Grainger Graphic

    Grainger

    Chicago, Illinois, United States

  • -

    Chicago, Illinois, United States

  • -

    Chicago, Illinois, United States

  • -

    Greater Chicago Area

  • -

    Nevada, United States

  • -

  • -

    Annaba, Algeria

  • -

  • -

    VHR Stallite image Classification

Education

  • University of Nevada-Las Vegas Graphic

    University of Nevada-Las Vegas

    -

  • -

    Visiting scholar at LISIC - Computer science, Signal and vision laboratory

  • -

    This Degree is equivalent to US-PhD degree, accredited by FIS International Organisation.

  • -

    This Degree is equivalent to US-MSc degree, accredited by FIS International Organisation

  • -

    This Degree is equivalent to US-BSc degree, accredited by FIS International Organisation

Licenses & Certifications

Publications

  • Fast CNN Surveillance Pipeline for Fine-Grained Vessel Classification and Detection in Maritime Scenarios

    AVSS 2016, IEEE

    Deep convolutional neural networks (CNNs) have
    proven very effective for many vision benchmarks in object
    detection and classification tasks. However, the computational
    complexity and object resolution requirements of
    CNNs limit their applicability in wide-view video surveillance
    settings where objects are small. This paper presents
    a CNN surveillance pipeline for vessel localization and
    classification in maritime video. The proposed pipeline
    is build upon the GPU…

    Deep convolutional neural networks (CNNs) have
    proven very effective for many vision benchmarks in object
    detection and classification tasks. However, the computational
    complexity and object resolution requirements of
    CNNs limit their applicability in wide-view video surveillance
    settings where objects are small. This paper presents
    a CNN surveillance pipeline for vessel localization and
    classification in maritime video. The proposed pipeline
    is build upon the GPU implementation of Fast-R-CNN
    with three main steps:(1) Vessel filtering and regions proposal
    using low-cost weak object detectors based on handengineered
    features. (2) Deep CNN features of the candidates
    regions are computed with one feed-forward pass
    from the high-level layer of a fine-tuned VGG16 network.
    (3) Fine-grained classification is performed using CNN features
    and a support vector machine classifier with linear
    kernel for object verification. The performance of the proposed
    pipeline is compared with other popular CNN architectures
    with respect to detection accuracy and evaluation
    speed. The proposed approach mAP of 61.10% was the
    comparable with Fast-R-CNN but with a 10× speed up (on
    the order of Faster-R-CNN) on the new Annapolis Maritime
    Surveillance Data-set.

    Other authors
    See publication

  • Off-the-Shelf CNN Features for Fine-Grained Classification of Vessels in a Maritime Environment

    Springer

    Convolutional Neural Networks (CNNs) have recently achie- ved spectacular performance on standard image classification benchmarks. Moreover, CNNs trained using large datasets such as ImageNet have performed effectively even on other recognition tasks and have been used as generic feature extraction tool for off-the-shelf classifiers. This paper, presents an experimental study to investigate the ability of off-the-shelf CNN features catch discriminative details of maritime vessels for…

    Convolutional Neural Networks (CNNs) have recently achie- ved spectacular performance on standard image classification benchmarks. Moreover, CNNs trained using large datasets such as ImageNet have performed effectively even on other recognition tasks and have been used as generic feature extraction tool for off-the-shelf classifiers. This paper, presents an experimental study to investigate the ability of off-the-shelf CNN features catch discriminative details of maritime vessels for fine-grained classification. An off-the-shelf classification scheme utilizing a linear support vector machine is applied to the high-level convolution features that come before fully connected layers in popular deep learning architectures. Extensive experimental evaluation compared OverFeat, GoogLeNet, VGG, and AlexNet architectures for feature extraction. Results showed that OverFeat features outperform the other architectures with a mAP = 0.7021 on the nine class fine-grained problem which was almost 0.02 better than its closest competitor, GoogLeNet, which performed best on smaller vessel types.

    Other authors
    See publication

  • Improved mean shift integrating texture and color features for robust real time object tracking

    The Visual Computer Journal, Springer

    The subject of mean shift algorithm for tracking the location of an object using a color model has recently gained considerable interest. However, the use of a color model to represent the tracked object is very sensitive to clutter interference, illumination changes, and the influence of background. Therefore, the applicability of basic color-based mean shift tracking is limited in many real world complex conditions. In this paper, we present a modified adaptive mean shift tracking algorithm…

    The subject of mean shift algorithm for tracking the location of an object using a color model has recently gained considerable interest. However, the use of a color model to represent the tracked object is very sensitive to clutter interference, illumination changes, and the influence of background. Therefore, the applicability of basic color-based mean shift tracking is limited in many real world complex conditions. In this paper, we present a modified adaptive mean shift tracking algorithm integrating a combination of texture and color features. We first suggest a new texture-based target representation based on spatial dependencies and co-occurrence distribution within interest target region for invariant target description, which is computed through so-scaled Haralick texture features. Then, to improve the tracking further, we propose an extension to the mean shift tracker where a combination of texture and color features are used as the target model. To be consistent to the scale change and complex non-rigid motions of the tracked target, we suggest to adapt the tracking window of the proposed algorithm with the real moving target mask at tracking over time. Many experimental results demonstrate the successful of target tracking using the proposed algorithm in many complex situations, where the basic mean shift tracker obviously fails. The performance of the proposed adaptive mean shift tracker is evaluated using the VISOR video Dataset, thermal infrared-acquired images sequences bench mark, and also some proprietary videos.

    See publication

  • An evolutionary scheme for decision tree construction

    Knowledge-Based Systems journal/ELSEVIER

    Classification is a central task in machine learning and data mining. Decision tree (DT) is one of the most popular learning models in data mining. The performance of a DT in a complex decision problem depends on the efficiency of its construction. However, obtaining the optimal DT is not a straightforward process. In this paper, we propose a new evolutionary meta-heuristic optimization based approach for identifying the best settings during the construction of a DT. We designed a genetic…

    Classification is a central task in machine learning and data mining. Decision tree (DT) is one of the most popular learning models in data mining. The performance of a DT in a complex decision problem depends on the efficiency of its construction. However, obtaining the optimal DT is not a straightforward process. In this paper, we propose a new evolutionary meta-heuristic optimization based approach for identifying the best settings during the construction of a DT. We designed a genetic algorithm coupled with a multi-task objective function to pull out the optimal DT with the best parameters. This objective function is based on three main factors: (1) Precision over the test samples, (2) Trust in the construction and validation of a DT using the smallest possible training set and the largest possible testing set, and (3) Simplicity in terms of the size of the generated candidate DT, and the used set of attributes. We extensively evaluate our approach on 13 benchmark datasets and a fault diagnosis dataset. The results show that it outperforms classical DT construction methods in terms of accuracy and simplicity. They also show that the proposed approach outperforms Ant-Tree-Miner (an evolutionary DT construction approach), Naive Bayes and Support Vector Machine in terms of accuracy and F-measure.

    See publication

Honors & Awards

  • Top 30 Leading Experts in AI

    MIT Technology Review Arabia

  • Timmy Awards, Best-Tech-Manager in Chicago

    Tech-in-Motion

Languages

  • Anglais

    Full professional proficiency

  • Français

    Full professional proficiency

  • Arabe

    Native or bilingual proficiency

  • Espagnol

    Elementary proficiency

Organizations

  • University Of Nevada Las Vegas, USA

    Research Scholar

    -

Recommendations received

2 people have recommended Fouad

Join now to view

More activity by Fouad

View Fouad’s full profile

  • See who you know in common
  • Get introduced
  • Contact Fouad directly
Join to view full profile

Other similar profiles

Explore more posts

Explore collaborative articles

We’re unlocking community knowledge in a new way. Experts add insights directly into each article, started with the help of AI.

Explore More

Add new skills with these courses

See all courses