Method for determining whether medication has been administered and server using same
Inventors
Assignees
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Publication Number
US-12310745-B2
Publication Date
2025-05-27
Expiration Date
Abstract
Provided is a server for determining whether medication has been administered, the server including: a transceiver receiving a video recorded by a wearable device; a memory storing a detection model and a confirmation model, wherein the detection model is trained to output whether each of preset targets appears in an image, and the confirmation model is trained to output whether medication has been administered, wherein the preset targets include an object related to a medicine or a medicine container and a posture related to medication administration; and one or more processors configured to detect the preset targets by inputting image frames of the video to the detection model and to determine whether medication has been administered by inputting confirmation model input data to the confirmation model, the confirmation model input data generated based on a detection result of the detection model.
Core Innovation
The invention relates to a medication-administration monitoring system that uses a wearable device to capture video including a plurality of image frames. A server receives the video and classifies the video into a specific category related to a specific object associated with a medication administration of a user, and based on the classified category determines the medication administration associated with the corresponding object of the user.
Each category-associated model includes both a detection model and a confirmation model. The detection model outputs a plurality of probability related to a plurality of objects and a plurality of postures at each of a plurality of time points corresponding to the plurality of image frames, and the confirmation model outputs whether medication corresponding to the classification has administered by a user based on receiving the plurality of probability related to the plurality of objects and the plurality of postures.
The disclosure further describes category-based approaches and alternatives, including using a classification model to select medication-administration type/category and applying category-specific monitoring, detection, and confirmation. It also describes a training pipeline with detection labels for object/posture and confirmation labels for administered vs non-administered, handling user variability and incomplete capture in training by using removed frames or sub-data, and computing medication adherence by comparing detected or confirmed medication administration with a medication administration schedule.
Claims Coverage
The independent claim coverage centers on wearable video category classification for medication administration objects and category-associated models that include detection and confirmation. Additional refinements include first-category versus second-category medication determination, model case selective identification of corresponding object and behavior, same-structure models trained on different object-specific data, and comparison to a medication administration schedule.
Wearable video category classification for medication administration objects
Receive video including a plurality of image frames from a wearable device, classify the video into a specific category related to a specific object associated with a medication administration of a user, and use the classification to route determination to a model associated with the category.
Category-associated model with detection and confirmation
Each model includes a detection model and a confirmation model, where the detection model outputs a plurality of probability related to a plurality of objects and a plurality of postures at each of a plurality of time points, and the confirmation model outputs whether medication corresponding to the classification has administered by a user based on receiving the plurality of probability.
First-category versus second-category medication determination
When the video is classified into a first category based on a first object, determine medication administration associated with the first object using a first model, and when classified into a second category based on a second object, determine medication administration associated with the second object using a second model.
Model case selective identification of corresponding object and behavior
Determine medication administration by identifying only the corresponding first object and its associated first behavior when using the first model, or only the corresponding second object and its associated second behavior when using the second model.
Same-structure models trained on different object-specific data
The first model and the second model share the same structure, with the first model trained using data associated with the first object and the second model trained using data associated with the second object.
Overall, the claims coverage focuses on category-based routing of medication-administration determination from wearable video, using models that combine detection of objects and postures over time with confirmation of whether medication corresponding to the classification has been administered, together with object-specific training and selective object or behavior handling.
Stated Advantages
Improved performance is demonstrated for approaches combining detection and confirmation, and for category and classification-based approaches.
Documented Applications
Medication adherence computation by comparing detected or confirmed medication administration information with a medication administration schedule.
Telemedicine reporting, including generating a secondary opinion report and a prescription report.
Interested in licensing this patent?