ACID Development Roadmap
The ACID project was initiated in September 2018 when Dr. Shih-Chung Kang joined the University of Alberta. At that time, the construction research community recognized the immense potential of deep learning in construction automation, yet it lacked a public image dataset. The original goal of ACID was to provide a public dataset to contribute to the community.
​
Under the supervision of Dr. Kang, Dr. Bo Xiao collected 10,000 construction images, developed object detection annotations for ACID, and conducted object detection algorithm analysis with the assistance of five undergraduate students from the University of Alberta within one year. In September 2019, Dr. Yiheng Wang joined the team, and the development of ACID image captioning annotations commenced, resulting in the annotation of 4,000 images with sentences by 2021. In January 2020, the research team published the ACID website and began sharing the ACID dataset with the community.
​
From 2021 to 2024, Dr. Bo Xiao led the research team and completed instance segmentation and image captioning annotations for all 10,000 images. In March 2024, Dr. Lei Hou and Dr. Hung-Lin Chi joined the ACID team, and the team finalized the algorithm analysis for instance segmentation and image captioning. By July 2024, the new ACID website had been published, making all three annotations of the 10,000 images available to the community!
CaRC Lab
Dr. Bo Xiao leads the Computing and Robotic Construction (CaRC) Lab, which is based in the Department of Civil, Environmental, and Geospatial Engineering at Michigan Technological University (MTU). Automation, digitalization, and robotics technologies are the key success factor for the fourth construction industry revolution (i.e. Construction 4.0), which can potentially enhance the efficiency, productivity, accuracy, and safety of the construction industry and address the complex problems surrounding our cities, environment, and the planet. The CaRC Lab has been and will continue to advance the digital transformation of the construction industry by adopting automated technologies (e.g., artificial intelligence (AI), robotics, and digital twinning).
CI Lab
Dr. Chi leads the Construction Informatics Laboratory (CI Lab), which is established aiming to advance construction-related computational technologies and applications through 1) enhancing the knowledge and
theoretical foundations in addressing construction education, planning, monitoring, and operational
challenges; 2) identifying best practices of optimal construction in terms of safety and productivity
and 3) matching emerging digital technologies to tackle practical construction problems. The
research directions for the lab team members include but are not limited to designing AI-enabled
new construction planning and optimization approaches, developing sensor-enabled construction
crew-machine interaction and user interface, integrating advanced information and management
systems for construction projects, and further, a particular focus lies in, establishing digital literacy
for construction engineering education using visualization technologies.
DSBL Lab
Dr. Hou leads the Digital and Sustainable Building Lab (DSBL) at RMIT University, situated within the Department of Civil and Infrastructure Engineering. The lab leverages state-of-the-art technologies including Virtual and Augmented Reality, advanced sensors, high-performance computing, and robotics to revolutionise building practices. Beyond supporting groundbreaking research, DSBL enriches the educational experience by integrating these technologies into RMIT's Smart Construction program. This initiative prepares students to tackle the intricate challenges posed by modern cities and environmental concerns through digital innovation.
AIRCon-Lab
Dr. Shih-Chung Kang leads the AIRCon-Lab, a AI and robotics lab that supports interdisciplinary research aims at introducing robotics to construction sites. Equipped with multiple robots, sensor networks, virtual reality devices, and computation units, the lab makes simulation possible for a variety of construction scenarios and develops innovative construction methods by integrating the sensors, artificial intelligence, and automatic machines. The robotics solutions propose significant productivity increases for construction projects, jobsite safety improvements, and reduced impacts due to labor shortages, especially in Northern America. To accomplish this goal, ACID is the essential data resource for the AIRCon-Lab uses to establish its artificial intelligence models.
Collaborate with ACID
We welcome collaboration with other researchers and enterprises. If you are interested in developing deep learning algorithms, extending the construction image dataset, or looking for solutions using artificial intelligence methods, we will be excited to work with you. Please contact Dr. Bo Xiao by email dr.boxiao@gmail.com if you would like to initiate a collaboration.