Deep ensemble learning method for zero-shot object detection

doi:10.1016/j.asoc.2025.114333

Deep ensemble learning method for zero-shot object detection

2025-11-26

0

PRE

AI

DOI：10.1016/j.asoc.2025.114333

Original

Original request for help

Save

Abstract

En 中文

• For the zero-shot learning (ZSD) model, we propose using the visual embedding space. This is because embedding features from the visual space into the semantic space can improve recognition. • With word vector representations, we propose to use the Random Multimodal Deep Learning Model (RMDL) to model textual descriptions using deep learning. • While our neural network model is highly adaptable and capable of learning the semantic space representation end-to-end, the focus of this paper is on establishing a highly effective multi-modality integration method. • The efficacy of our proposed method on benchmark datasets such as MSCOCO, Visual Genome, and ILSVRC 2012, among others, has been demonstrated through extensive experiments.

Deep ensemble learning method for zero-shot object detection

Abstract

Journal

Organization