Cleanroom Personnel Dress Code Compliance Detection Dataset

Currently, with the rapid development of high-tech manufacturing industries, cleanrooms have become the core of production in industries such as semiconductors and biopharmaceuticals. However, problems of dress code non-compliance due to human factors still exist, affecting the safety of production and the quality of products. Existing manual inspection methods have issues with low efficiency and strong subjectivity, necessitating a technical means of automatically identifying dress code compliance. This dataset focuses on solving the problem of wear compliance detection in image recognition, meeting the needs of automated and standardized management in cleanroom environments. Data collection was conducted in cleanrooms of multiple industry-leading companies, using high-definition camera equipment to gather a large number of samples at different times and from various angles. Stringent quality control measures such as multiple rounds of labeling, consistency checks, and expert reviews were introduced, with annotation completed by a professional team with relevant industry background, comprising more than 20 people. The data underwent standardized cropping, color balancing, and image denoising in the preprocessing stage and is ultimately stored in structured JPG format grouped by individuals, facilitating rapid search and processing. The core advantage of the dataset lies in high labeling accuracy and consistency, with an error labeling rate of less than 1%. New automated labeling tools and data augmentation techniques, such as flipping and rotation, were introduced to enhance the robustness of model recognition. This dataset has been successfully applied in actual production, helping enterprises significantly reduce human inspection time and improve dress compliance rates by over 30%. Compared to similar datasets, the cleanroom personnel dress code compliance detection dataset has a significant improvement in sample diversity and resolution, covering content from general environments to specific industries, with great scarcity and versatility, leading the industry's data standards and application standards.