Image Categorization Using Scene-Context Scale Based on Random Forests

Yousun KANG; Hiroshi NAGAHASHI; Akihiro SUGIMOTO

doi:10.1587/transinf.E94.D.1809

Image Categorization Using Scene-Context Scale Based on Random Forests

Yousun KANG, Hiroshi NAGAHASHI, Akihiro SUGIMOTO

Full Text Views

0

Cite this

Summary :

Scene-context plays an important role in scene analysis and object recognition. Among various sources of scene-context, we focus on scene-context scale, which means the effective scale of local context to classify an image pixel in a scene. This paper presents random forests based image categorization using the scene-context scale. The proposed method uses random forests, which are ensembles of randomized decision trees. Since the random forests are extremely fast in both training and testing, it is possible to perform classification, clustering and regression in real time. We train multi-scale texton forests which efficiently provide both a hierarchical clustering into semantic textons and local classification in various scale levels. The scene-context scale can be estimated by the entropy of the leaf node in the multi-scale texton forests. For image categorization, we combine the classified category distributions in each scale and the estimated scene-context scale. We evaluate on the MSRC21 segmentation dataset and find that the use of the scene-context scale improves image categorization performance. Our results have outperformed the state-of-the-art in image categorization accuracy.

Publication: IEICE TRANSACTIONS on Information Vol.E94-D No.9 pp.1809-1816

Publication Date: 2011/09/01

Publicized

Online ISSN: 1745-1361

DOI: 10.1587/transinf.E94.D.1809

Type of Manuscript: PAPER

Category: Image Recognition, Computer Vision

Cite this

Copy

Yousun KANG, Hiroshi NAGAHASHI, Akihiro SUGIMOTO, "Image Categorization Using Scene-Context Scale Based on Random Forests" in IEICE TRANSACTIONS on Information, vol. E94-D, no. 9, pp. 1809-1816, September 2011, doi: 10.1587/transinf.E94.D.1809.
Abstract: Scene-context plays an important role in scene analysis and object recognition. Among various sources of scene-context, we focus on scene-context scale, which means the effective scale of local context to classify an image pixel in a scene. This paper presents random forests based image categorization using the scene-context scale. The proposed method uses random forests, which are ensembles of randomized decision trees. Since the random forests are extremely fast in both training and testing, it is possible to perform classification, clustering and regression in real time. We train multi-scale texton forests which efficiently provide both a hierarchical clustering into semantic textons and local classification in various scale levels. The scene-context scale can be estimated by the entropy of the leaf node in the multi-scale texton forests. For image categorization, we combine the classified category distributions in each scale and the estimated scene-context scale. We evaluate on the MSRC21 segmentation dataset and find that the use of the scene-context scale improves image categorization performance. Our results have outperformed the state-of-the-art in image categorization accuracy.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.E94.D.1809/_p

Copy

@ARTICLE{e94-d_9_1809,
author={Yousun KANG, Hiroshi NAGAHASHI, Akihiro SUGIMOTO, },
journal={IEICE TRANSACTIONS on Information},
title={Image Categorization Using Scene-Context Scale Based on Random Forests},
year={2011},
volume={E94-D},
number={9},
pages={1809-1816},
abstract={Scene-context plays an important role in scene analysis and object recognition. Among various sources of scene-context, we focus on scene-context scale, which means the effective scale of local context to classify an image pixel in a scene. This paper presents random forests based image categorization using the scene-context scale. The proposed method uses random forests, which are ensembles of randomized decision trees. Since the random forests are extremely fast in both training and testing, it is possible to perform classification, clustering and regression in real time. We train multi-scale texton forests which efficiently provide both a hierarchical clustering into semantic textons and local classification in various scale levels. The scene-context scale can be estimated by the entropy of the leaf node in the multi-scale texton forests. For image categorization, we combine the classified category distributions in each scale and the estimated scene-context scale. We evaluate on the MSRC21 segmentation dataset and find that the use of the scene-context scale improves image categorization performance. Our results have outperformed the state-of-the-art in image categorization accuracy.},
keywords={},
doi={10.1587/transinf.E94.D.1809},
ISSN={1745-1361},
month={September},}

Copy

TY - JOUR
TI - Image Categorization Using Scene-Context Scale Based on Random Forests
T2 - IEICE TRANSACTIONS on Information
SP - 1809
EP - 1816
AU - Yousun KANG
AU - Hiroshi NAGAHASHI
AU - Akihiro SUGIMOTO
PY - 2011
DO - 10.1587/transinf.E94.D.1809
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E94-D
IS - 9
JA - IEICE TRANSACTIONS on Information
Y1 - September 2011
AB - Scene-context plays an important role in scene analysis and object recognition. Among various sources of scene-context, we focus on scene-context scale, which means the effective scale of local context to classify an image pixel in a scene. This paper presents random forests based image categorization using the scene-context scale. The proposed method uses random forests, which are ensembles of randomized decision trees. Since the random forests are extremely fast in both training and testing, it is possible to perform classification, clustering and regression in real time. We train multi-scale texton forests which efficiently provide both a hierarchical clustering into semantic textons and local classification in various scale levels. The scene-context scale can be estimated by the entropy of the leaf node in the multi-scale texton forests. For image categorization, we combine the classified category distributions in each scale and the estimated scene-context scale. We evaluate on the MSRC21 segmentation dataset and find that the use of the scene-context scale improves image categorization performance. Our results have outperformed the state-of-the-art in image categorization accuracy.
ER -