The task of person re-identification (ReID) is to match images of the same person over
multiple non-overlapping camera views. Due to the variations in visual factors, previous …

Recent methods for deep metric learning have been focusing on designing different
contrastive loss functions between positive and negative pairs of samples so that the …

Person re-identification is the process of recognizing an individual across multiple camera
views. It is essential for an extensive range of applications related to security and biometrics …

In unsupervised learning of image features without labels, especially on datasets with fine-
grained object classes, it is often very difficult to tell if a given image belongs to one specific …

Existing long-tailed recognition methods focus on learning global image representation by
re-weighing, re-sampling, or global representation learning. However, we observe that …

It has been long recognized that deep neural networks are sensitive to changes in spatial
configurations or scene structures. Image augmentations, such as random translation …

Operational risk (OR) is usually caused by losses due to human errors, inadequate or
defective internal processes, system failures, or external events that affect an organization …

Person re-identification (ReID) aims to match and track people in a surveillance system with
non-overlapping camera views. It is a key challenge for person ReID to learn robust and …

Aprendizado Incremental lida com um aumento contınuo no número de classes ao longo do
tempo. O principal problema nesse contexto é o” esquecimento catastrófico”, em que os …

Whilst convolutional neural networks (CNN) offer state-of-the-art performance for
classification and detection tasks in computer vision, their successful adoption in defence …