Correlation-based ConvNet for Small Object Detection in Videos
The detection of small objects is of particular
interest in many real applications. In this paper, we propose
STDnet-ST, a novel approach to small object detection in video
using spatial information operating alongside temporal video
information. STDnet-ST is an end-to-end spatio-temporal convolutional neural network that detects small objects over time
and correlates pairs of the top-ranked regions with the highest
likelihood of containing small objects. This architecture links the
small objects across the time as tubelets, being able to dismiss
unprofitable object links in order to provide high-quality tubelets.
STDnet-ST achieves state-of-the-art results for small objects on
the publicly available USC-GRAD-STDdb and UAVDT video
datasets.
keywords:
Publication: Congress
1624015059138
June 18, 2021
/research/publications/correlation-based-convnet-for-small-object-detection-in-videos
The detection of small objects is of particular
interest in many real applications. In this paper, we propose
STDnet-ST, a novel approach to small object detection in video
using spatial information operating alongside temporal video
information. STDnet-ST is an end-to-end spatio-temporal convolutional neural network that detects small objects over time
and correlates pairs of the top-ranked regions with the highest
likelihood of containing small objects. This architecture links the
small objects across the time as tubelets, being able to dismiss
unprofitable object links in order to provide high-quality tubelets.
STDnet-ST achieves state-of-the-art results for small objects on
the publicly available USC-GRAD-STDdb and UAVDT video
datasets. - B. Bosquet, M. Mucientes and V. Brea
publications_en