Sheet Feature Detection Using Overlaid Drawing Pages and ML
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The AEC industry faces challenges in automatically detecting sheet titles and numbers due to the lack of a standard format, leading to manual and time-consuming processes, especially when dealing with varying formats of drawing sheets.
Innovation Solution
A computer-implemented method using a machine learning model to detect target objects on a document page by generating an image from overlapping drawing sheets, identifying cells within the target area, extracting informational features, and training the model to recognize sheet titles and numbers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If manual methods are used to detect sheet titles and numbers, then detection accuracy can be maintained through human judgment, but productivity is reduced due to time-consuming manual processing
Solution Approach 1:
The patent replaces manual mechanical detection with an automated machine learning-based detection system. The system uses trained models to automatically identify sheet titles and numbers in drawings, eliminating the need for manual human judgment and significantly improving detection speed and productivity.
Solution Approach 2:
The detection system is designed to autonomously process drawings without human intervention. The machine learning model automatically detects and extracts sheet titles and numbers from various drawing formats, enabling the system to serve itself in performing the detection task that previously required manual human effort.
2Productivity
If automatic detection methods are implemented, then productivity is improved through automated processing, but detection reliability deteriorates due to format variations across different drawings
Solution Approach 1:
The patent employs multiple machine learning models with different parameters and architectures to handle various drawing formats. By changing and adjusting model parameters based on the specific drawing format being processed, the system maintains high detection reliability across diverse formats while preserving automated processing efficiency.
Solution Approach 2:
The detection system is designed with universal machine learning models that can handle multiple drawing formats and styles. The models are trained on diverse datasets encompassing various formats, enabling them to universally detect sheet titles and numbers across different drawing types without requiring format-specific manual processing.
3Adaptability or versatility
If machine learning models are trained on diverse drawing formats, then adaptability is improved to handle various formats, but device complexity increases due to the training and deployment of multiple models
Solution Approach 1:
The patent performs preliminary training of machine learning models on diverse drawing formats during the development phase. By pre-training models on comprehensive datasets that include various formats, styles, and conventions, the system achieves high adaptability to different formats while reducing the complexity of deploying multiple specialized models, as the pre-trained models can handle diverse inputs with minimal additional configuration.
Data Source
AI summary
A computer-implemented method for detecting a target object on a document page that includes detecting a sample target area on a sample document page, generating an image by overlapping a plurality of sample document pages with one another, and detecting one or more cells within the sample target area on the image. The sample target area includes a sample target object. The method further includes extracting one or more informational features from each of the one or more cells. The one or more informational features define characteristics of a corresponding cell of the one or more cells. A machine learning model is trained using the one or more informational features extracted from each of the one or more cells, to detect the sample target object. A target object on a document page is detected using the trained machine learning model.