Automated Object Marking Transfer in Sensor Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The generation of image and video annotations for machine learning is costly and often requires human intervention, as automating this task is limited or not possible, leading to high annotation times, such as over an hour for semantic segmentation.
Innovation Solution
A method and system for object marking in sensor data that detects a scene in multiple states, allowing for the partial acceptance and automation of object markings across different data sets using sensors like cameras and LIDAR, with AI modules and neural networks to recognize and transform annotations, reducing the need for complete reannotation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If human operators perform object marking manually, then annotation accuracy and reliability are improved, but time consumption and cost increase significantly
Solution Approach 1:
The system performs preliminary object marking in a first state, then re-detects the same scene in a second state and accepts the preliminary marking for the second state as well. This preliminary action in the first state benefits both states, reducing the need for complete re-annotation and significantly cutting time consumption while maintaining reliability through the initial human-performed accurate marking.
2Reliability
If complete reannotation is performed for each data set, then annotation accuracy is maintained, but cost and effort increase
Solution Approach 1:
The system creates a copy of the object marking from the first data set and applies it to the second data set after re-detection. Instead of performing complete reannotation, the system copies the preliminary marking and accepts it for the second state, significantly reducing annotation effort while maintaining accuracy through the copying of reliable markings.
3Productivity
If automation is increased for object marking, then productivity improves, but measurement precision and reliability deteriorate
Solution Approach 1:
The system enables self-service by allowing the preliminary object marking to serve dual purposes for both the first and second states. The marking process serves itself across different states through re-detection and acceptance, improving productivity without sacrificing precision because the same reliable marking is reused rather than replaced by potentially less accurate automated markings.
Data Source
AI summary
A method and a system for improved object marking in sensor data, as the result of which an at least partially automated annotation of objects or object classes in a recorded data set is possible. The method provides that a scene is detected in a first state by at least one sensor. An association of a first object marking with at least one object contained in the scene in a first data set containing the scene in the first state then takes place. The similar or matching scene is subsequently detected in a second state that is different from the first state by the at least one sensor, and an at least partial acceptance of the first object marking, contained in the first data set, for the object recognized in the second state of the scene as a second object marking in a second data set takes place.