Vehicle Object Recognition Relearning With User-Guided Adjustment Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing software programs, particularly those used for object recognition processing, struggle to be updated with adjustment content according to user requests, as the required parameter adjustments are unclear, leading to suboptimal recognition accuracy and user satisfaction.
Innovation Solution
An information processing system that includes a vehicle with a camera and recognition section, a data logger, and a server, which collects user-requested parameters, relearns the recognition section, and updates the software to align with user preferences, using adjustment images to guide parameter adjustments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the SW is updated with adjustment content according to user requests, then user satisfaction and recognition accuracy are improved, but it is difficult to determine how much adjustment should be performed on what kind of parameter to achieve the desired adjustment content
Solution Approach 1:
The patent introduces an adjustment image as an intermediary tool that visually represents the relationship between parameter values and recognition results. This adjustment image serves as a mediator between the user's desired adjustment content and the actual parameter settings, making it easy to understand how much adjustment should be performed on each parameter without requiring complex technical knowledge
Solution Approach 2:
The system implements feedback by displaying adjustment images that show the predicted recognition results based on current parameter settings. Users can see the effect of parameter adjustments in real-time and make informed decisions about what adjustments to make, creating a closed-loop system where users can iteratively refine parameters based on visual feedback
2Adaptability or versatility
If the parameter is adjusted to change the adjustment content (such as sensitivity or intensity), then the SW can be updated to meet user requests, but it requires clear understanding of parameter relationships which is currently unavailable
Solution Approach 1:
The adjustment image acts as an intermediary that translates complex parameter relationships into visual information. Instead of requiring users to directly understand and manipulate multiple parameters, the adjustment image presents the effects of parameter changes in an intuitive visual format, making the system adaptable to user requests while remaining easy to operate
Solution Approach 2:
The system enables easy parameter adjustment by displaying adjustment images that show the relationship between parameter values and recognition outcomes. Users can easily change parameters based on the visual feedback from adjustment images, allowing flexible adaptation to user requests without requiring deep technical understanding of parameter interactions
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The present disclosure relates to an information processing apparatus, an information processing method, an information processing system, and a program that enable an SW to be easily updated with an adjustment content according to a request of a user. Vehicle accumulation information including an image for relearning the SW that functions as an object recognition section that recognizes an object around a vehicle and an object recognition result is acquired from the image around the vehicle, adjustment content requested by the user of the vehicle regarding the relearning of the SW is decided, and the SW is relearned on the basis of the vehicle accumulation information and the adjustment content. The present disclosure can be applied to an automated driving technology.