a fully-labeled image dataset to advance indoor objects detection
A fully-labeled image dataset provides a unique resource for reproducible research inquiries and data analyses in several computational fields, such as computer vision, machine learning and deep learning machine intelligence. With the present contribution, a large-scale fully-labeled image dataset is provided, and made publicly and freely available to the research community. The current dataset entitled MCIndoor20000 includes more than 20,000 digital images from three different indoor object categories, including doors, stairs, and hospital signs. To make a comprehensive dataset regarding current challenges exist in indoor objects modeling, we cover a multiple set of variations in images, such as rotation, intra-class variation plus various noise models. The current dataset is freely and publicly available for any academic, educational, and research purposes.
Value of the Data
The MCIndoor20000 dataset collected in Marshfield Clinic (https://www.marshfieldclinic.org/), and it presents various digital images of three guideline indoor objects, including clinic signs, doors and stairs.
Computational vision have become ubiquitous in our daily life, with a variety of applications ranging from face recognition and fingerspelling to surveillance systems and healthcare informatics. Core to many of these applications is image classification and recognition which is defined as an automatic task that assigns a label from a fixed set of categories to an input image. The MCIndoor20000 dataset is a resource for use by the computer vision and deep learning community, and it boosts image classification research.
To provide a comprehensive image classification repository, the current dataset covers several object model variations involved from the perspectives of computer vision and deep learning strategies. The variations include viewpoint variation, intra-class variation, rotation, noisy conditions (e.g., Gaussian, Poisson), and occlusion.
The MCIndoor20000 dataset assists reproducible research and allows rapid application development (RAD) and fast prototyping to the deep learning and computer vision community.
Sample images from the categories covered by the MCIndoor20000
Links to the data
Gaussian Noise Model (1)
Gaussian Noise Model (2)
Poisson Noise Model
Salt and Pepper Noise Model
- Fereshteh S. Bashiri
- Eric LaRose
- Peggy Peissig
- Ahmad P. Tafti
We wish to thank Marshfield Clinic which allowed us to collect the MCIndoor20000 dataset. Our special thanks goes to Daniel Wall and Anne Nikolai at Marshfield Clinic Research Institute (https://www.marshfieldresearch.org/). We appreciate their great help and contributions in collecting the dataset and preparing the current paper. Fereshteh S. Bashiri would like to thank the Summer Research Internship Program (SRIP) (http://www.marshfieldresearch.org/srip) at Marshfield Clinic Research Institute (MCRI) for financial support.
The MCIndoor20000 dataset is fully explained in the following paper. Any publication using the database would encourage to reference to:
 Bashiri, F.S., LaRose, E., Peissig, P. and Tafti, A.P., 2018. MCIndoor20000: a fully-labeled image dataset to advance indoor objects detection. Data in Brief. [Paper]