3rd Continual Learning Workshop Challenge on Egocentric Category and Instance Level Object Understanding

Pellegrini, Lorenzo; Zhu, Chenchen; Xiao, Fanyi; Yan, Zhicheng; Carta, Antonio; Lange, Matthias De; Lomonaco, Vincenzo; Sumbaly, Roshan; Rodríguez, Paul; Vázquez, David

doi:10.48550/arxiv.2212.06833

Cited by 2 publications

(15 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…EgoObjects was introduced during the CLVision CVPR22 workshop [19]. It depicts ego-centric videos taken indoor environments focusing on everyday objects.…”

Section: Related Workmentioning

confidence: 99%

“…To investigate the impact of NR on training and model performance in OOD scenarios, we conducted an experiment on the EgoObjects dataset [19].…”

Section: Experiments On Egoobjectsmentioning

confidence: 99%

“…The EgoObjects dataset was initially introduced in the CLVision challenge at CVPR 2022 [19] to create a benchmark for evaluating continual object detectors. In this challenge scenario, the EgoObjects videos were divided into five distinct tasks based on the labels of the focused objects within each video.…”

Section: Experiments On Egoobjectsmentioning

confidence: 99%

“…First of all, we conduct an extensive experimental study on the EgoObjects dataset [19], which are thoughtfully designed to exhibit varying degrees of NR across different object classes. Our investigations aim to reveal the influence of NR on model evaluation, particularly its effect on the forgetting rate.…”

Section: Introductionmentioning

confidence: 99%

“…EgOAK has been specifically designed to tackle these inquiries by offering a comprehensive evaluation framework that considers the dynamics of NR and domain changes. The core idea behind EgOAK is the alternating training on tasks from two datasets, EgoObjects [19] and OAK [28]. It ensures that the model is exposed to tasks of different domains in a controlled manner.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Comparative Study of Natural Replay and Experience Replay in Online Object Detection

Wagner,

Pellerin,

Huet

2023

2023 IEEE/CVF International Conference on Computer Vision Workshops (ICCVW)

View full text Add to dashboard Cite

Online Object Detection (OOD) algorithms play a crucial role in dynamic and real-world computer vision applications. In these scenarios, models are trained on a data stream where old class samples are revisited, a phenomenon known as Natural Replay (NR). During training, NR occurs unevenly across object categories, leading to evaluation metrics biased towards the most frequently revisited classes. Existing benchmarks lack proper quantification of NR and depict short-term training scenarios on a single domain. As a result, evaluating generalization capabilities and forgetting rates of models become challenging in OOD. In this paper, we address the challenges surrounding the evaluation of OOD models by proposing two key contributions. Firstly, we define a metric to quantify NR in an OOD scenario and show how NR is related to class specific forgetting. Secondly, we introduce a novel benchmark, EgOAK, which introduces a long-term training scenario that involves frequent domain shifts. It allows the evaluation of models' generalization capabilities and forgetting of knowledge on past domains. Our results in this OOD setting reveal that Experience Replay, a memory-based method, is particularly effective for better generalization to new domains and for preserving past knowledge. Leveraging replay from memory helps to address the low natural replay rate for rarely revisited classes, resulting in improved adaptability and reliability of models in dynamic environments.

show abstract

“…EgoObjects was introduced during the CLVision CVPR22 workshop [19]. It depicts ego-centric videos taken indoor environments focusing on everyday objects.…”

Section: Related Workmentioning

confidence: 99%

“…To investigate the impact of NR on training and model performance in OOD scenarios, we conducted an experiment on the EgoObjects dataset [19].…”

Section: Experiments On Egoobjectsmentioning

confidence: 99%

Section: Experiments On Egoobjectsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Comparative Study of Natural Replay and Experience Replay in Online Object Detection

Wagner,

Pellerin,

Huet

2023

2023 IEEE/CVF International Conference on Computer Vision Workshops (ICCVW)

View full text Add to dashboard Cite

show abstract

Continual Object Detection with Deep Neural Networks

Menezes

View full text Add to dashboard Cite

The rapid technological development in the past decades has significantly increased the amount of available data in the world. Naturally, models that scale with the size of the available data, such as Deep Neural Networks, have become the primary strategy for several research fields with abundant data (e.g., computer vision and natural language processing). With this large data availability, research on learning models that can adapt incrementally to continual streams of data has been encouraged. In this way, the field of Continual Learning proposes to study the ability to learn consecutive tasks without losing performance on the previously trained ones. In computer vision, researchers have mainly focused their efforts on incremental classification tasks, but continual object detection also deserves attention due to its vast range of applications in robotics and autonomous vehicles. In fact, this scenario is even more complex than conventional classification, given the occurrence of instances of classes that are unknown at the time but can appear in subsequent tasks as a new class to be learned, resulting in missing annotations and conflicts with the background label. Since this field is in its early stages, research in continual object detection still offers several opportunities and lacks methodology conventions. This Ph.D. thesis investigates the field more thoroughly and identifies possible links with related areas such as general continual learning and neural network pruning. Specifically, we proposed the first systematic review on the topic, developed two metrics for improving the analysis of performance in incremental detection scenarios, investigated which exemplar selection method works best for replay-based continual detection strategies, and explored different ways to identify and penalize important task parameters across sequential updates. To validate our proposals and claims, we conducted experiments and reported results comparable to the current state-of-the-art in popular detection benchmarks (i.e., PASCAL VOC) adapted to the incremental setting, as well as in real-world datasets and applications. The findings presented in this thesis were also put into practice in two applications. Firstly, they were tested in the 3 rd CLVISION Challenge, where we were able to achieve the 3 rd place in the continual instance detection track. Secondly, they were applied to the continual aerial inspection of transmission towers at TAESA, the largest Brazilian electric power transmission company, to improve the automation of their inspection pipeline.

show abstract

3rd Continual Learning Workshop Challenge on Egocentric Category and Instance Level Object Understanding

Cited by 2 publications

References 0 publications

Comparative Study of Natural Replay and Experience Replay in Online Object Detection

Comparative Study of Natural Replay and Experience Replay in Online Object Detection

Continual Object Detection with Deep Neural Networks

Contact Info

Product

Resources

About