We developed an AI component that translates a medical image from one domain to another using AI Style Transfer Technique. The goal of the component is to enable re-use of existing software components and machine learning models that are employed to process a specific type of images i.e. images of tissues stained with one type of reagent across all other reagents.
How Artificial Intelligence can help with Cancer Image Translation | AI in Medical Imaging
Summary
Challenge
The challenge was to develop a system that could autonomously detect and assess the cleanliness of workplace areas, recognizing objects and distinguishing clean from dirty or messy areas. Additionally, the project aimed to detect people within these areas while respecting privacy concerns, ensuring that images with people were not processed or stored.
Solution by AI Superior
In the first phase, AI Superior developed a robust object detection model capable of identifying and classifying various workplace objects (including bins, cleaning products, food waste, cups, glasses, backpacks, handbags, and more) in zenithal photos. The model also categorized areas as clean or messy based on dynamic rules.
The second phase focused on developing a model to detect people within workplace areas. Precision was crucial to avoid processing or storing images containing identifiable individuals, respecting privacy concerns. AI Superior implemented a person detection model that could identify partial representations of individuals, such as a foot or hand, without capturing recognizable facial features.
Outcome and Implications
Despite limited publicly available zenithal photo datasets, AI Superior successfully developed a highly precise model for identifying objects within workplace environments and distinguishing clean from dirty or messy areas. This system’s ability to detect and classify objects provided actionable insights, facilitating timely cleaning interventions while the privacy-focused person detection feature ensured responsible and ethical handling of sensitive information.
English 
German 
Arabic 
French 
Dutch 
Spanish