In this paper, we explore a method for self-supervised learning, tackling the challeng- ing tasks of visual object detection and retrieval. To this end, we ...

Our evaluation is based on a set of ten objects with manual ground truth annotation in almost 5000 frames extracted from instructional videos from the web. We ...

In this paper, we explore a method for self-supervised learning, tackling the challeng- ing tasks of visual object detection and retrieval. To this end, we ...

Nov 12, 2024 · In this work, we propose to exploit the natural correlation in narrations and the visual presence of objects in video, to learn an object ...

In this paper we address a different problem of self-supervised object detection and retrieval from unlabeled and unconstrained videos. The closest work to our ...

Reimplementation of the paper "Learning to Detect and Retrieve Objects from Unlabeled Videos" [https://arxiv.org/abs/1905.11137]. Using the dataset provided by ...

... Instructional videos are a natural source of self-supervision, as they contain both speech and visual information. For instance, Amrani et al. [2] recently ...

Dive into the research topics of 'Self-supervised object detection and retrieval using unlabeled videos'. Together they form a unique fingerprint. Sort by ...

Dive into the research topics of 'Self-supervised object detection and retrieval using unlabeled videos'. Together they form a unique fingerprint. Sort by ...

Dive into the research topics of 'Self-supervised object detection and retrieval using unlabeled videos'. Together they form a unique fingerprint. Sort by ...