Recurrent Convolutional Fusion for RGB-D Object Recognition
- 6 June 2019
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Robotics and Automation Letters
- Vol. 4 (3) , 2878-2885
Abstract
Providing robots with the ability to recognize objects like humans has always been one of the primary goals of robot vision. The introduction of RGB-D cameras has paved the way for a significant leap forward in this direction thanks to the rich information provided by these sensors. However, the robot vision community still lacks an effective method to synergically use the RGB and depth data to improve object recognition. In order to take a step in this direction, we introduce a novel end-to-end architecture for RGB-D object recognition called recurrent convolutional fusion (RCFusion). Our method generates compact and highly discriminative multi-modal features by combining RGB and depth information representing different levels of abstraction. Extensive experiments on two popular datasets, RGB-D Object Dataset and JHUIT-50, show that RCFusion significantly outperforms state-of-the-art approaches in both the object categorization and instance recognition tasks. In addition, experiments on the more challenging Object Clutter Indoor Dataset confirm the validity of our method in the presence of clutter and occlusion. The code is publicly available at: “https://github.com/MRLoghmani/rcfusion”.Keywords
All Related Versions
This publication has 22 references indexed in Scilit:
- Large-Margin Multi-Modal Deep Learning for RGB-D Object RecognitionIEEE Transactions on Multimedia, 2015
- Multimodal deep learning for robust RGB-D object recognition2015
- Hypercolumns for object segmentation and fine-grained localization2015
- Learning Phrase Representations using RNN Encoder–Decoder for Statistical Machine Translation2014
- Histogram of Oriented Normal Vectors for Object Recognition with a Depth Sensor2013
- A learned feature descriptor for object recognition in RGB-D data2012
- A large-scale hierarchical multi-view RGB-D object dataset2011
- Improving the Fisher Kernel for Large-Scale Image Classification2010
- Long Short-Term MemoryNeural Computation, 1997
- Sequential neural text compressionIEEE Transactions on Neural Networks, 1996