Convolutional hypercube pyramid for accurate RGB-D object category and instance recognition
Deep learning based methods have achieved unprecedented success in solving several computer vision problems involving RGB images. However, this level of success is yet to be seen on RGB-D images owing to two major challenges in this domain: training data deficiency and multi-modality input dissimila...
| Main Authors: | , , |
|---|---|
| Format: | Proceeding Paper |
| Language: | English English |
| Published: |
IEEE
2016
|
| Subjects: | |
| Online Access: | http://irep.iium.edu.my/60177/ http://irep.iium.edu.my/60177/3/60177%20Convolutional%20Hypercube%20Pyramid.pdf http://irep.iium.edu.my/60177/2/60177%20Convolutional%20Hypercube%20Pyramid.scopus.pdf |
| _version_ | 1848785445688705024 |
|---|---|
| author | Mohd Zaki, Hasan Firdaus Shafait, Faisal Mian, Ajmal |
| author_facet | Mohd Zaki, Hasan Firdaus Shafait, Faisal Mian, Ajmal |
| author_sort | Mohd Zaki, Hasan Firdaus |
| building | IIUM Repository |
| collection | Online Access |
| description | Deep learning based methods have achieved unprecedented success in solving several computer vision problems involving RGB images. However, this level of success is yet to be seen on RGB-D images owing to two major challenges in this domain: training data deficiency and multi-modality input dissimilarity. We present an RGB-D object recognition framework that addresses these two key challenges by effectively embedding depth and point cloud data into the RGB domain. We employ a convolutional neural network (CNN) pre-trained on RGB data as a feature extractor for both color and depth channels and propose a rich coarse-to-fine feature representation scheme, coined Hypercube Pyramid, that is able to capture discriminatory information at different levels of detail. Finally, we present a novel fusion scheme to combine the Hypercube Pyramid features with the activations of fully connected neurons to construct a compact representation prior to classification. By employing Extreme Learning Machines (ELM) as non-linear classifiers, we show that the proposed method outperforms ten state-of-the-art algorithms for several tasks in terms of recognition accuracy on the benchmark Washington RGB-D and 2D3D object datasets by a large margin (upto 50% reduction in error rate). |
| first_indexed | 2025-11-14T16:53:16Z |
| format | Proceeding Paper |
| id | iium-60177 |
| institution | International Islamic University Malaysia |
| institution_category | Local University |
| language | English English |
| last_indexed | 2025-11-14T16:53:16Z |
| publishDate | 2016 |
| publisher | IEEE |
| recordtype | eprints |
| repository_type | Digital Repository |
| spelling | iium-601772018-08-06T08:13:23Z http://irep.iium.edu.my/60177/ Convolutional hypercube pyramid for accurate RGB-D object category and instance recognition Mohd Zaki, Hasan Firdaus Shafait, Faisal Mian, Ajmal QA75 Electronic computers. Computer science Deep learning based methods have achieved unprecedented success in solving several computer vision problems involving RGB images. However, this level of success is yet to be seen on RGB-D images owing to two major challenges in this domain: training data deficiency and multi-modality input dissimilarity. We present an RGB-D object recognition framework that addresses these two key challenges by effectively embedding depth and point cloud data into the RGB domain. We employ a convolutional neural network (CNN) pre-trained on RGB data as a feature extractor for both color and depth channels and propose a rich coarse-to-fine feature representation scheme, coined Hypercube Pyramid, that is able to capture discriminatory information at different levels of detail. Finally, we present a novel fusion scheme to combine the Hypercube Pyramid features with the activations of fully connected neurons to construct a compact representation prior to classification. By employing Extreme Learning Machines (ELM) as non-linear classifiers, we show that the proposed method outperforms ten state-of-the-art algorithms for several tasks in terms of recognition accuracy on the benchmark Washington RGB-D and 2D3D object datasets by a large margin (upto 50% reduction in error rate). IEEE 2016 Proceeding Paper PeerReviewed application/pdf en http://irep.iium.edu.my/60177/3/60177%20Convolutional%20Hypercube%20Pyramid.pdf application/pdf en http://irep.iium.edu.my/60177/2/60177%20Convolutional%20Hypercube%20Pyramid.scopus.pdf Mohd Zaki, Hasan Firdaus and Shafait, Faisal and Mian, Ajmal (2016) Convolutional hypercube pyramid for accurate RGB-D object category and instance recognition. In: 2016 IEEE International Conference on Robotics and Automation (ICRA), 16-21 May 2016, Stockholm, Sweden. https://ieeexplore.ieee.org/document/7487310/ 10.1109/ICRA.2016.7487310 |
| spellingShingle | QA75 Electronic computers. Computer science Mohd Zaki, Hasan Firdaus Shafait, Faisal Mian, Ajmal Convolutional hypercube pyramid for accurate RGB-D object category and instance recognition |
| title | Convolutional hypercube pyramid for accurate RGB-D object category and instance recognition |
| title_full | Convolutional hypercube pyramid for accurate RGB-D object category and instance recognition |
| title_fullStr | Convolutional hypercube pyramid for accurate RGB-D object category and instance recognition |
| title_full_unstemmed | Convolutional hypercube pyramid for accurate RGB-D object category and instance recognition |
| title_short | Convolutional hypercube pyramid for accurate RGB-D object category and instance recognition |
| title_sort | convolutional hypercube pyramid for accurate rgb-d object category and instance recognition |
| topic | QA75 Electronic computers. Computer science |
| url | http://irep.iium.edu.my/60177/ http://irep.iium.edu.my/60177/ http://irep.iium.edu.my/60177/ http://irep.iium.edu.my/60177/3/60177%20Convolutional%20Hypercube%20Pyramid.pdf http://irep.iium.edu.my/60177/2/60177%20Convolutional%20Hypercube%20Pyramid.scopus.pdf |