DPDenis Pellerin

Denis Pellerin

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Denis Pellerin has a total of 7 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Studying Forgetting in Faster R-CNN for Online Object Detection: Analysis Scenarios, Localization in the Architecture, and Mitigation

    by Baptiste Wagner, Denis Pellerin, Sylvain Huet

    Online Object Detection (OOD) requires learning new object categories from a stream of images, similar to an agent exploring new environments. In this context, the widely used Faster R-CNN (Region Convolutional Neural Network) architecture faces catastrophic forgetting: the acquisition of new knowledge leads to the loss of previously learned information. In this paper, we investigate the learning and forgetting mechanisms of the Faster R-CNN in OOD through three main contributions. First, we observe that the forgetting curves of the Faster R-CNN exhibit patterns similar to those described in human memory studies by Hermann Ebbinghaus: knowledge is lost exponentially over time and recall improves knowledge retention. Second, we present a new methodology for analysing the Faster R-CNN architecture and quantifying forgetting across the Faster R-CNN components. We show that forgetting is mainly localised in the Softmax classification layer. Finally, we propose a new training strategy for OOD called Configurable Recall (CR). CR performs recalls on old data using images stored in a memory buffer with variable frequency and recall length to ensure efficient learning. CR also masks the logits of old objects in the softmax classification layer to mitigate forgetting. We evaluate our strategy against state-of-the-art methods on three OOD benchmarks. We analyse the effectiveness of different recall types in mitigating forgetting and show that CR outperforms existing methods.


  • Open Access

    SHAPE-BASED INDIVIDUAL/GROUP DETECTION FOR SPORT VIDEOS CATEGORIZATION

    by Costas Panagiotakis, Emmanuel Ramasso, Georgios Tziritas, Michèle Rombaut, Denis Pellerin

    We present a shape-based method for automatic people detection and counting without any assumption concerning camera motion. In order to evaluate the robustness of the proposed method, we apply it for classifying athletics videos into two classes: videos of individual and videos of team sports. The videos used are real and characterized by dynamic and unconstrained environment. Moreover, in the case of team sport, we propose a shape deformations based method for running/hurdling discrimination (activity recognition). Robust, adaptive and independent from color, illumination changes and the camera motion, the proposed features are combined in the Transferable Belief Model (TBM) framework providing a two-level (frames and shot) video categorization. Experimental results of 97% of accuracy for individual/team sport categorization using a dataset of 252 real videos of athletic meetings, acquired by moving cameras under varying view angles, indicate the stability and the good performance of the proposed scheme.


  • Open Access

    State Filtering and Change Detection Using TBM Conflict Application to Human Action Recognition in Athletics Videos

    by Emmanuel Ramasso, Michle Rombaut, Denis Pellerin

    In this paper, we propose a tool called temporal credal filter with conflict-based model change (TCF-CMC) to smooth belief functions online in transferable belief model (TBM) framework. The TCF-CMC takes temporal aspects of belief functions into account and relies on conflict information explicitly modelled in TBM when combining beliefs. TBM fusion, in addition to uncertainty, takes into account imprecision and conflict inherent to features. The TCF-CMC takes part in a wider system for human action recognition in videos. The whole system is tested on 62 videos (11000 images) with moving camera and real conditions where the TCF-CMC improves running, jumping, falling and standing-up actions recognition in high jump, pole vault, long jump and triple jump activities. The TCF-CMC is also compared to hidden Markov models. Lastly, a TBM rules-based modelling is compared to Gaussian mixture.