Techniques for extracting machine-readable information from P and IDs
Inventors
GARDNER, MARC-ANDRÉ • Jahjah, Karl-Alexandre
Assignees
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Publication Number
US-12175337-B2
Publication Date
2024-12-24
Expiration Date
Abstract
In example embodiments, techniques are provided for using machine learning to extract machine-readable labels for text boxes and symbols in P&IDs in image-only formats. A P&ID data extraction application uses an optical character recognition (OCR) algorithm to predict labels for text boxes in a P&ID. The P&ID data extraction application uses a first machine learning algorithm to detect symbols in the P&ID and return a predicted bounding box and predicted class of equipment for each symbol. One or more of the predicted bounding boxes may be decimate by non-maximum suppression to avoid overlapping detections. The P&ID data extraction application uses a second machine learning algorithm to infer properties for each detected symbol having a remaining predicted bounding box. The P&ID data extraction application stores the predicted bounding box and a label including the predicted class of equipment and inferred properties in a machine-readable format.
Core Innovation
The invention addresses extracting machine-readable labels for text boxes and symbols in piping and instrumentation drawings (P&IDs) provided in image-only formats. It predicts, using optical character recognition (OCR), bounding boxes and text for text boxes in a P&ID image-only format, and the OCR uses a domain-specific dictionary in language modeling based on knowledge of words and codes used in P&IDs in a given industry or based on a particular list of words and codes known to be used by a specific user in P&IDs.
In parallel, the invention detects symbols in the P&ID using a first machine learning algorithm and returns a predicted bounding box and a predicted class of equipment for each detected symbol. To avoid overlapping detections, it decimates one or more of the predicted bounding boxes, and for each detected symbol having a remaining predicted bounding box, it infers properties using a second machine learning algorithm.
Finally, the invention stores, in a machine-readable format, the predicted bounding box and a label for each detected symbol having a remaining predicted bounding box. The label includes the predicted class of equipment and the inferred properties, and the machine-readable format is configured to store the predicted results in a memory of a computing device.
Claims Coverage
The document includes three independent claims (a method, a computing device, and a non-transitory computing device readable medium). Across these independent claims, the inventive features focus on OCR for P&ID text boxes using domain-specific language modeling, machine-learning detection and classification of equipment symbols, decimating overlapping symbol detections, inferring symbol properties, and storing predicted bounding boxes and labels in a machine-readable format.
OCR with domain-specific dictionary in language modeling for P&ID text boxes
Predicting, using optical character recognition (OCR), bounding boxes and text for text boxes in a piping and instrumentation drawing (P&ID) in an image-only format, wherein the OCR uses a domain-specific dictionary in language modeling based on knowledge of words and codes used in P&IDs in a given industry or based on a particular list of words and codes known to be used by a specific user in P&IDs.
Machine-learning symbol detection returning equipment class
Detecting, using machine learning, symbols in the P&ID and returning a predicted bounding box and a predicted class of equipment for each symbol.
Decimating predicted bounding boxes to avoid overlapping detections
Decimating one or more of the predicted bounding boxes to avoid overlapping detections.
Inferring properties for detected symbols with remaining bounding boxes
Inferring, using machine learning, properties for each detected symbol having a remaining predicted bounding box.
Storing machine-readable labels including equipment class and inferred properties
Storing, in a machine-readable format, for each detected symbol having a remaining predicted bounding box, the predicted bounding box and a label including the predicted class of equipment and inferred properties.
Across the independent claims, the core coverage is a pipeline that combines OCR of image-only P&ID text boxes using a domain-specific dictionary in language modeling, machine-learning detection and classification of equipment symbols, decimation of overlapping symbol bounding boxes, machine-learning property inference for remaining symbols, and storage of predicted bounding boxes and machine-readable labels including equipment class and inferred properties.
Stated Advantages
Not explicitly described in patent.
Documented Applications
Not explicitly described in patent.
Interested in licensing this patent?