[Data Annotation Case Study] SMRT (Singapore)

Images from monitoring the conditions of railway tracks


SMRT is a major public transport company in Singapore with its core business of managing and operating train services on major lines in Singapore.

Currently in SMRT, there is a system to monitor the conditions of the tracks using high-resolution cameras mounted on the MRT trains. SMRT has been exploring ways to increase the accuracy of this system so that they can improve the productivity and efficiency of the image reviewing process. SMRT has trialled ABEJA solution to help build up an algorithm to interface with the camera system.

SMRT is keen to introduce machine learning software to monitor the conditions of their track.
ABEJA platform provides a data annotation service that SMRT’s engineers could utilise in order to develop a labelled dataset to train an algorithm to detect defects. Through the ABEJA platform, image data would be uploaded into the cloud and displayed for engineers to label the parts that are identified as defects. Due to its code-free interface, annotators only need to mark a bounding box over an identified anomaly and choose the correct customised label.

The detection of defects from uploaded images data taken by cameras is a laborious task as some defects are not obvious and require subject matter experts with years of experience to identify.  There are also multiple defects to identify and categorise.

With the ABEJA platform, the ease of data annotation will support Machine Learning and Deep Learning adoption, which will also reduce the effort needed for track condition monitoring.

■ About ABEJA
With its corporate philosophy “Implement a Fruitful World”, ABEJA pursues industrial structure reform by the power of state-of-the-art technology particularly AI. Since establishment in 2012, we develop Deep Learning technology in-house and provide AI solutions to society. We provide end-to-end support for AI implementation and utilization across industries such as retail, manufacturing, logistics, infrastructure, etc. Having established a local subsidiary in Singapore in March 2017, ABEJA continues to expand its operations globally, focusing on ASEAN countries especially.

Company:ABEJA, Inc.
Representative: Yousuke Okada, President & Representative Director, CEO
Location: 10F NBF Platinum Tower, 1-17-3 Shirokane, Minato-ku, Tokyo
Description of business: A provider of AI solutions to society leveraging Deep Learning technology
URL :https://abejainc.com

《 Inquiries 》
ABEJA, Inc., Globalsales
Email: globalsales@abejainc.com

Back to News list