BY MENGYING ZHANG, IOWA STATE UNIVERSITY; LINSEY GRIFFIN, UNIVERSITY OF MINNESOTA; RUI LI, IOWA STATE UNIVERSITY; AND GUOWEN SONG, IOWA STATE UNIVERSITY
Properly fitting protective gloves are crucial for the safety and efficiency of firefighters. Ill-fitting gloves can hinder performance, reduce dexterity, and even lead to injuries. According to the National Fire Protection Association (NFPA), between 2016 and 2020, 40% of fireground injuries involved extremities, with 10 to 12% affecting hands or fingers.1 Furthermore, research from the National Institute for Occupational Safety and Health (NIOSH) indicates that glove fit issues affect up to 30% of male and 62% of female firefighters.2 Poor glove fit can compromise thermal protection, blood circulation, and grip, which are all vital during firefighting operations. In extreme cases, firefighters may even remove gloves to perform certain tasks, putting themselves at significant risk.
Current glove sizing systems are outdated, relying on measurements derived from military personnel that do not adequately represent the diverse hand dimensions of modern firefighters.2 Moreover, traditional methods like manual measurements or 3D scanning are inefficient for large-scale data collection of firefighters’ hand dimensions. Addressing these challenges requires innovative data collection methods and a reevaluation of glove design to better reflect the needs of both male and female firefighters.
Challenges in Glove Fit and Design
The importance of proper glove fit for firefighters is widely acknowledged, yet several key challenges continue to hinder meaningful progress in this area. One of the most significant obstacles is the lack of comprehensive anthropometric data specific to firefighters. Currently, there is no robust database of hand measurements tailored to the firefighting population. As a result, glove designs often fall short of meeting the unique needs of firefighters. This issue is particularly pronounced for female firefighters, whose gloves are frequently derived from male hand dimensions. Such designs fail to account for differences in hand proportions or the positions of critical points like joints, leading to poor fit and reduced functionality. Simply downsizing a male glove pattern does not address these disparities, leaving female firefighters with gloves that hinder rather than support their performance.
Another critical aspect of glove fit is its dynamic nature. A proper fit is not solely about matching static measurements but also about how well the glove accommodates hand movements and postures during real-world tasks. Firefighters rely on gloves that provide both security and flexibility under demanding conditions. Feedback from firefighters about their hand movements during various tasks is essential to creating designs that enhance performance without compromising comfort or protection.
Addressing these issues also requires stronger collaboration among stakeholders. Researchers, glove manufacturers, and fire fighters must work together to tackle fit and functionality problems effectively. Without direct input from those who use the gloves daily, key design features like joint pleats, knuckle protectors, and grip pads may fail to perform as intended. By fostering collaboration and incorporating user feedback into the design process, stakeholders can ensure that gloves meet the practical needs of firefighters while advancing safety and performance.
Issues in Current Turnout Gloves
The NFPA glove sizing system was designed with a foundation rooted in anthropometric data from military personnel. On paper, it features a standardized size chart with regular intervals and linearity for key dimensions like index finger length and hand breadth. However, in practice, the reality is far more complex and challenging. Across different brands, significant inconsistencies in glove sizing have been reported, creating frustration for firefighters trying to find gloves that fit well-even when adhering to NFPA guidelines.
To understand the extent of this problem, researchers at the Human Dimensioning Laboratory at the University of Minnesota evaluated the consistency and reliability of NFPA-compliant firefighter gloves’ sizing systems. The full size range of five NFPA-compliant commercial glove brands (one type/model of glove for each brand) were measured to evaluate the linearity of the sizing system and the consistency of quality (Figure A1).
The key dimensions, index finder length, and hand breadth were used to evaluate the brands’ sizing systems and applied the same interval rules dictated by the NFPA for length and width dimensions. The results showed that none of the five brands showed linearity across their sizing system, and they did not follow a similar pattern to the NFPA. Significant overlap in sizes was seen across the brands as well as inconsistent intervals demonstrating the need for improved quality. The visuals of the brands’ sizing system based on measurements from the glove exterior show significant overlap between sizes, which would make it very difficult for firefighters to select the appropriate size glove. The analysis reveals that, across all dimensions, glove styles, and brands, the sizing systems lack both linearity and consistency. The finding that glove measurements do not exhibit linearity is consistent with previous research by Sokolowski & Griffin.3 Consequently, the current five brands of firefighting gloves have ambiguous size intervals and do not align with NFPA recommendations, indicating either a failure in adhering to the NFPA system for glove design or significant deficiencies in manufacturing and quality control tolerances.
To understand the dimensional consistency and quality across the five glove brands, the right and left gloves (across all sizes commercially available from each brand) were measured. A total of 24 measurement items were extracted from the gloves, and the measurements were categorized according to type: digit lengths, digit breadths and depths, digit circumferences, and measurements in the palm/dorsum region).
All five brands showed statistically significant differences in digit length and major inconsistencies for other measurement groups. Inconsistency in finger length sizing might be responsible for fit-related issues highlighted in previous research. Using tolerance levels based on the NFPA sizing system of +/- 3mm for length measurements, +/- 2mm for width measurements, and +/- 4mm for the palm region, the researchers compared the left-right differences by brand and tolerance range. The results show that as little as 38% of glove pairs fell within the tolerance range for digit length, 49% for digit breadth/depth, 26% for digit circumference, and 36% for the palm/dorsum measurements. The maximum difference observed between the left and right paired gloves across all measurements was as much as 17mm. The largest difference in digit length observed between the left and right paired gloves was 13.5mm. Many of the brands left-right paired glove difference exceeded more than twice the tolerance levels.
Identifying the dimensional differences of each firefighter glove and comparing them with established tolerance levels is important for the following reasons: (1) This research highlights the importance of maintaining consistent dimensions before considering other design factors that may impact fit. Size consistency is a basic precondition for a good fit. (2) This research draws attention to size inspection and tolerance settings that manufacturers need to improve. (3) This research provides a foundation for quantitative reliability verification when manufacturers develop glove systems. In particular, if manufacturers and researchers embark on innovating glove design, material development, or sizing systems without understanding the inherent manufacturing errors in the glove, problems with sizing/fit will continue. In this regard, maintaining quality and consistent sizes within tolerance levels will improve fit and will also improve the impact of future innovation.
In addition to evaluating the consistency and reliability of gloves, the researchers at the University of Minnesota have developed a method to compare the interior and exterior of multi-layered structural firefighter gloves to assist in understanding the hand-glove relationship, as well as providing insights for future design innovation. Through innovative CT scanning application, the researchers are able to measure the air gap between the interior glove (that the hand interacts with) and the exterior glove. Figure A2 shows the results of this, demonstrating the effectiveness of using advanced technology to quantify the fit and performance of multi-layered protective gloves.
AI-Assisted Glove Sizing for Firefighters
To tackle these issues, recent research at Iowa State University in collaboration with University of Minnesota and Texas A&M University-Corpus Christi has centered on developing an AI-assisted approach to firefighter glove Sizing and Analysis for Fit Enhancement and improved HAND protection (AI-SAFEHAND). Supported by a Department of Homeland Security Assistance to Firefighters Grant, this project led to the creation of an AI-powered hand anthropometric tool to assist the revision of the current NFPA glove sizing system for firefighters. The tool includes an Al-based algorithm that accurately detects hand key points and automatically calculates hand dimensions. Additionally, a mobile application (APP) was developed, enabling firefighters to capture 2D photos of their hand’s dorsal and thumb sides, using a standard letter-sized paper as a reference background. The general workflow of the mobile APP-based hand anthropometric tool is illustrated in Figure A3. And so far, we have made great progress on this project.
AI-Driven Hand Anthropometry
The AI-SAFEHAND project introduces an innovative approach to measuring firefighter hand dimensions using a mobile phone application, available for both Android and iOS platforms. This cutting-edge tool allows firefighters to capture images of their hands by placing them against a standard-sized reference background, such as a letter-sized paper. These images are then analyzed using advanced Al techniques, offering a comprehensive solution to understanding hand dimensions with precision.
The process begins with image preprocessing, where the captured images are resized to preserve aspect ratios. A Fast Fourier Transform (FFT) is applied to extract critical features while reducing noise, ensuring clean and accurate data. Next, hand masking and inpainting techniques are applied using powerful models like the Segment Anything Model (SAM) and LaMa. These tools generate precise hand masks and refine the background, ensuring the hand outline is clear and free from interference.
With the processed images, key point detection is performed using the YOLOv8 model. This step identifies crucial points on the hand, such as joints and edges, and enhances accuracy through additional edge point detection and midpoint calculations. Finally, measurement calculation is carried out using the Euclidean distance formula to determine the distances between key points, providing detailed hand dimensions. The accuracy of these measurements is rigorously evaluated using metrics like Mean Squared Error and Root Mean Squared Error.
This Al-driven process enables a deeper understanding of the diverse hand shapes and sizes among firefighters. Preliminary results have shown a substantial reduction in measurement errors, highlighting the potential of this approach to provide a more accurate and reliable foundation for designing gloves that fit better and enhance firefighter performance. By leveraging the AI-SAFEHAND project, the firefighting community is taking a significant step toward improved safety and functionality in PPE.
Next Steps
The AI-SAFEHAND project is paving the way for a revolutionary approach to addressing glove fit challenges faced by firefighters. By harnessing the power of Al and mobile technology, this initiative offers a comprehensive, data-driven solution to glove sizing that accommodates the diverse needs of both male and female firefighters. Through close collaboration between researchers, manufacturers, and firefighters, the project aims to ensure gloves are not only well-fitting but also capable of performing effectively in the demanding conditions firefighters face daily.
Future iterations of the AI-SAFEHAND system promise even greater advancements. Planned features include tailored recommendations, providing firefighters with personalized glove suggestions based on their unique hand dimensions and preferences. A 3D hand visualization tool will allow users to view a 3D-rendered model of their hand, enhancing understanding and interaction with the sizing system. Additionally, an enhanced feedback system will enable firefighters to share their experiences with specific glove models, contributing valuable insights to improve glove design and performance for the entire community.
As part of its development journey, the AI-SAFEHAND project will hold a preliminary public test of the app at FDIC 2025. Attending firefighters are invited to try out the app and provide feedback, helping to shape the future of this groundbreaking system.
The AI-SAFEHAND initiative is a critical step forward in solving long-standing issues with glove fit. With its innovative approach and commitment to ongoing improvement, this project has the potential to significantly enhance firefighter safety, efficiency, and overall job performance. By integrating advanced technology with direct user input, AI-SAFEHAND is setting a new standard for personal protective equipment in the firefighting industry.
REFERENCES
- Campbell, Richard. “Firefighter Injuries on the Fireground.” NFPA, 1 Aug. 2022, bit.ly/3D4qJYt.
- Hsiao, Hongwei, et al. “Firefighter Hand Anthropometry and Structural Glove Sizing: A New Perspective.” Human Factors, vol. 57, no. 8, 2015, pp. 1359-1377, bit.ly/49v9yLO.
- Sokolowski, Susan, and Linsey Griffin. “Method to Develop a Better Performance Glove Pattern Block Using 3d Hand Anthropometry.” Proceedings of the Human Factors and Ergonomics Society Annual Meeting, vol. 64, no. 1, Dec. 2020, pp. 1008-1012, bit.ly/4g6qJWD.
