Tampere University of Technology

TUTCRIS Research Portal

Faster Bounding Box Annotation for Object Detection in Indoor Scenes

Research output: Chapter in Book/Report/Conference proceedingConference contributionScientificpeer-review

Standard

Faster Bounding Box Annotation for Object Detection in Indoor Scenes. / Adhikari, Bishwo; Peltomäki, Jukka; Puura, Jussi; Huttunen, Heikki.

2018 7th European Workshop on Visual Information Processing (EUVIP). IEEE, 2018.

Research output: Chapter in Book/Report/Conference proceedingConference contributionScientificpeer-review

Harvard

Adhikari, B, Peltomäki, J, Puura, J & Huttunen, H 2018, Faster Bounding Box Annotation for Object Detection in Indoor Scenes. in 2018 7th European Workshop on Visual Information Processing (EUVIP). IEEE, European Workshop on Visual Information Processing, 1/01/00. https://doi.org/10.1109/EUVIP.2018.8611732

APA

Vancouver

Author

Adhikari, Bishwo ; Peltomäki, Jukka ; Puura, Jussi ; Huttunen, Heikki. / Faster Bounding Box Annotation for Object Detection in Indoor Scenes. 2018 7th European Workshop on Visual Information Processing (EUVIP). IEEE, 2018.

Bibtex - Download

@inproceedings{f9444cdf5b494251a1fa80ffd87e9973,
title = "Faster Bounding Box Annotation for Object Detection in Indoor Scenes",
abstract = "This paper proposes an approach for rapid bounding box annotation for object detection datasets. The procedure consists of two stages: The first step is to annotate a part of the dataset manually, and the second step proposes annotations for the remaining samples using a model trained with the first stage annotations. We experimentally study which first/second stage split minimizes to total workload. In addition, we introduce a new fully labeled object detection dataset collected from indoor scenes. Compared to other indoor datasets, our collection has more class categories, diverse backgrounds, lighting conditions, occlusions and high intra-class differences. We train deep learning based object detectors with a number of state-of-the-art models and compare them in terms of speed and accuracy. The fully annotated dataset is released freely available for the research community.",
author = "Bishwo Adhikari and Jukka Peltom{\"a}ki and Jussi Puura and Heikki Huttunen",
year = "2018",
month = "11",
doi = "10.1109/EUVIP.2018.8611732",
language = "English",
isbn = "978-1-5386-6898-6",
publisher = "IEEE",
booktitle = "2018 7th European Workshop on Visual Information Processing (EUVIP)",

}

RIS (suitable for import to EndNote) - Download

TY - GEN

T1 - Faster Bounding Box Annotation for Object Detection in Indoor Scenes

AU - Adhikari, Bishwo

AU - Peltomäki, Jukka

AU - Puura, Jussi

AU - Huttunen, Heikki

PY - 2018/11

Y1 - 2018/11

N2 - This paper proposes an approach for rapid bounding box annotation for object detection datasets. The procedure consists of two stages: The first step is to annotate a part of the dataset manually, and the second step proposes annotations for the remaining samples using a model trained with the first stage annotations. We experimentally study which first/second stage split minimizes to total workload. In addition, we introduce a new fully labeled object detection dataset collected from indoor scenes. Compared to other indoor datasets, our collection has more class categories, diverse backgrounds, lighting conditions, occlusions and high intra-class differences. We train deep learning based object detectors with a number of state-of-the-art models and compare them in terms of speed and accuracy. The fully annotated dataset is released freely available for the research community.

AB - This paper proposes an approach for rapid bounding box annotation for object detection datasets. The procedure consists of two stages: The first step is to annotate a part of the dataset manually, and the second step proposes annotations for the remaining samples using a model trained with the first stage annotations. We experimentally study which first/second stage split minimizes to total workload. In addition, we introduce a new fully labeled object detection dataset collected from indoor scenes. Compared to other indoor datasets, our collection has more class categories, diverse backgrounds, lighting conditions, occlusions and high intra-class differences. We train deep learning based object detectors with a number of state-of-the-art models and compare them in terms of speed and accuracy. The fully annotated dataset is released freely available for the research community.

U2 - 10.1109/EUVIP.2018.8611732

DO - 10.1109/EUVIP.2018.8611732

M3 - Conference contribution

SN - 978-1-5386-6898-6

BT - 2018 7th European Workshop on Visual Information Processing (EUVIP)

PB - IEEE

ER -