DESC:FIELD OF THE INVENTION
The present invention relates to optical character recognition. More particularly, the present disclosure relates to systems and methods for optical character recognition (OCR) of handwritten text or digitized text.
The method further includes processing the text using optical character recognition (OCR) engine and predicting/correlating the prescribed text with medicine via QR code, Barcode, image recognition and pharma invoice.
BACKGROUND OF THE INVENTION
Optical character recognition (OCR) is a technology that converts scanned documents, images, or PDF files into editable and searchable text. OCR software works by analyzing the pixels in an image or document and identifying the characters based on their shape and patterns.
There are many OCR tools available that can be used to extract text from prescriptions, including online tools and standalone software programs. Some OCR tools are specifically designed to recognize medical documents and may have features that are tailored to the formatting and language used in prescription documents.
A medical prescription is scanned, or the document that is converted into an electronic format, such as PDF or image file and a processor performing OCR is used to extract text from the medical prescription. The extracted text can be edited and saved as a new document, or it can be used to search for specific information within the document.
It is important to note that the accuracy of OCR can vary depending on the quality of the original document and the capabilities of the OCR tool being used. It may be necessary to manually review and correct any errors that are introduced during the OCR process.
An OCR prediction model is a machine learning model that is trained to recognize and classify text in images or documents. In the context of medicinal prescriptions, an OCR prediction model could be used to extract and classify different types of information from prescription documents, such as the patient's name, the prescription date, the drug name, and the dosage instructions.
To create an OCR prediction model for doctors’ prescriptions, a large dataset of prescription documents needs to be stored, and the relevant information in each document is labeled. Further, this labeled dataset is used to train a machine learning model to recognize and classify the different types of information in the prescriptions.
Once, the OCR prediction model is trained, it can be used to process new prescription documents and extract the relevant information from them. The model can be fine-tuned and improved over time by continuing to train it on additional labeled data.
Building an OCR prediction model requires a strong understanding of machine learning concepts and techniques, as well as experience working with large datasets and building machine learning models. It may be necessary to seek assistance from a machine learning expert or data scientist.
Several potential challenges and problems can arise when using OCR technology in the pharmacy setting. These include:
Handwriting: Interpretation of hand-written prescriptions preferred by doctors can be a significant problem for data analysis. Despite directions by governing bodies asking doctors to prescribe the names of the drugs in capital letters for better legibility, the doctors have not complied with the guidelines. Standard handwriting is still not maintained by doctors. The pharmacists leverage their domain knowledge and awareness of the brand preferred by local doctors based on prescription flow to interpret the prescriptions. Whenever required the pharmacist may also speak to the prescribing doctor for clarification.
Accuracy: OCR technology can sometimes introduce errors when extracting text from documents, especially if the original document is of low quality or has formatting that is difficult for the OCR software to interpret. These errors can lead to incorrect or incomplete information being extracted from the prescription, which can have serious consequences for the patient’s safety.
Formatting: Pharmacy prescriptions often have specific formatting requirements, such as the use of certain fonts and layout elements. OCR software may not always be able to accurately interpret these formatting elements, which can lead to incorrect or incomplete information being extracted.
Complex language: Medical terminology and jargon can be complex and may be difficult for OCR software to accurately recognize and interpret. This can lead to errors or misunderstandings when extracting information from prescription documents.
Integration with other systems: OCR technology may need to be integrated with other systems, such as electronic health records or pharmacy management systems, in order to be used effectively in the pharmacy setting. This can require additional work and resources to set up and maintain.
Overall, it is important to carefully consider the potential challenges and limitations of OCR technology when using it in the pharmacy setting and to take steps to mitigate these issues as much as possible. This may involve implementing quality checks and error-correction measures, as well as training staff on how to use the OCR technology correctly.
Hence, there is a need to provide a customized OCR-based system that can be localized in nature and is used to extract accurate information in a handwritten or digitized text and to provide a link between medicinal prescription and medicine via QR code, barcode, image recognition, and pharma invoice.
OBJECT OF THE INVENTION
The primary object of the present invention is to provide a system used for continuous digital aggregation of real-time granular data from fragmented doctor’s prescriptions collected from pharmacies using a combination of optical character recognition (OCR) and artificial intelligence (AI) interpretation.
Another object of the present invention is to provide an OCR-based technology for linkage between medicinal prescription and medicine via QR code, Barcode, image recognition, and pharma invoice.
A further object of the present invention is to provide a localized data collector-based OCR technology used for categorizing different medicine data types.
SUMMARY OF THE INVENTION
This invention enables a system for optical character recognition (OCR). More particularly, the present disclosure relates to systems and methods for optical character recognition (OCR) of handwritten or digitized text.
Embodiments of the present disclosure describe systems and methods for extracting symbols from a digitized object for character recognition.
Embodiments of the present disclosure relate to a system to provide automatic handwriting character recognition using optical character recognition (OCR).
This invention enables a system for optical character recognition (OCR) and a readable storage medium that bridges the gap between medicinal prescription and pharmaceutical medicines. The system uses a combination of OCR and AI to convert handwritten or printed prescriptions into machine-encoded text in real-time. This enables the identification of doctors based on their prescription designs and content.
Recognition of handwritten text is far more challenging than machine-generated text, which can include printed text on the invoice/bill, QR code, or barcode on a medicinal product, because of the virtually infinite ways a human can write in different formats and styles. Also, a fixed standard has not been maintained by all medicinal professionals.
The invention consists of a system based on Optical Character Recognition (OCR) which acts as a prediction model for pharmacy prescriptions, a large dataset of doctor’s prescriptions is collected by using several technologies that can be used to detect and identify medicine packages using a smartphone, these technologies can include a scanned image of a doctor’s prescription, medicinal product image, detection of a QR code or Barcode, scanned copy of a pharmacy bill or invoice.
The above-obtained prescription data is further processed by an OCR engine which results in a set of predicted data, which is compared with the medicine datasets obtained via different modes including QR code, Barcode, image recognition and Pharma invoice.
DETAILED DESCRIPTION OF DRAWINGS
To understand the invention and to see how it may be carried out in practice, embodiments will now be described, by way of non-limiting examples only, with reference to the accompanying drawings
Figure 1 illustrates a flow chart depicting a set of prescription data being processed by an OCR engine which results in a set of predicted data which is compared with the medicine datasets obtained via different modes including QR code, Barcode, image recognition and pharma invoice.
Figure 2 illustrates a table depicting the possible linkage between different predicted data items and datasets, showing the possible matches with respective items, for example, item 1 matches 99% with item 4, and item 2 matches 68% with 4.
DETAILED DESCRIPTION OF THE INVENTION
A localized optical character recognition (OCR) based system is herein described with numerous specific details to provide a complete understanding of the invention. However, these specific details are exemplary and should not be treated as a limitation to the scope of the invention.
Throughout this specification, the word “comprises” or variations such as “comprises or comprising”, will be understood to imply the inclusions of a stated element, integer or step, or group of elements, integers, or steps, but not the exclusions of any other element, integer or step or group of elements, integers, or steps.
The present invention features a system and method for automatically linking the prescription with the medicinal product. The method used is a combination system of Artificial Intelligence (AI) and optical character recognition (OCR) to convert the typed, handwritten, or printed prescriptions electrically or mechanically into machine-encoded text in real time to identify doctors by their prescription design and content.
The invention comprises several modules, including a data acquisition module, OCR engine, AI model, branding database, processing module, and linkage module. The data acquisition module collects prescription data from various sources, such as scanned images of prescriptions, product image recognition, QR code and barcode detection, and pharmacy invoices. The OCR engine processes the data to generate machine-encoded text. The AI model categorizes medicinal drug groups based on the processed prescription data with a predetermined threshold of similarities. The branding database maps medicine brands that can be co-prescribed.
The invention features a system for providing a scanned image of a prescription to be processed, the scanned image is obtained by a processor performing optical character recognition (OCR) on the prescription, further with an Artificial Intelligence model performing medicinal drug groups with a certain threshold of similarities, wherein each medicinal group corresponds to similar content, branding database to map the brands that can be co-prescribed.
The processing module acquires data from a scanned image of a prescription, the image of a medicinal product, details detection from a QR code or Barcode, and medicinal description details from a Pharmacy invoice.
The above-obtained prescription data is further processed by an OCR engine which results in a set of predicted data, which is compared with the medicine datasets obtained via different modes including QR code, Barcode, image recognition and Pharma invoice.
The processing module is further segregated into a branding database, handwriting database, prescription details, acute and chronic use of medication, and co-prescribed medications.
The above-obtained information is used by a pharmacist to develop an association build of medicinal products by creating a linkage between prescription and pharmaceutical medicine.
The following example is put forth to provide those of ordinary skill in the art with a complete disclosure and description of how the methods and systems claimed herein are performed and evaluated and are intended to be purely exemplary of the invention and are not intended to limit the scope of what the inventor regards as their invention.
EXAMPLE
There are several technologies that can be used to detect and identify medicine packages using a smartphone.
One example is a scanned image of a Prescription, which involves a smartphone’s camera to take a clear picture of the Doctor’s prescription, the details of a doctor’s prescription may carry a multitude of invaluable data including the Doctor’s qualification and specialty, clinic address, prescription date, patient’s information – name, age, gender, clinical history and symptoms, Diagnosis, drug name, dose, duration of therapy, all these details are scanned and recognized by an OCR.
The second example is Product image recognition, which involves a smartphone’s camera taking a clear picture of the product’s package and running this image through OCR software it works by analyzing the pixels in an image and by extracting the name and other details of the medicines.
The third example involves detecting information from a QR code a smartphone QR / barcode scanner can be used to scan the code and get full details of the medicinal product including the name of the medicine, contents of the drug, Active Pharmaceutical Ingredient (API), manufacturing details including MRP, Manufacturing date, expiry date of the product, an accurate and up to date information can be obtained.
There are also apps available that can be used to identify medicine packages using these technologies. Some of these apps allow you to scan the barcode or take a picture of the medicine package and will provide the name and other details about the medicine in response. Other apps may require you to enter the medicine's name or other identifying information manually.
The fourth example involves detecting information from a Pharmacy Invoice / Bill, this includes a description of the product, manufacturing details, batch no, expiry date, rate, the quantity of the medicine, and amount.
All the above-mentioned technologies can be used to detect and identify medicinal details.
An OCR software may not be accurately able to interpret all the medicinal / prescription details, which can lead to incorrect or incomplete information being extracted. Hence, OCR technology may need to be integrated with other systems, such as the mentioned above like Pharmacy management bills, QR codes, barcodes, and Prescription details. These systems can be used by an operator / Pharmacist to link the captured data with the doctor’s prescription eventually leading to a better accuracy rate.
The above obtained OCR data is further compared and linked by the pharmacist/chemist; comparison of the above data obtained in 1 can be linked with 2 or 3 or 4 or a combination thereof.
It's important to note that the accuracy of these technologies can vary depending on the quality of the image, the capabilities of the machine learning model, and other factors. It may be necessary to manually verify the information provided by the app or technology to ensure that it is accurate and up to date. ,CLAIMS:1. A data processing system for monitoring pharmaceutical prescriptions at various points in dispensing of prescription, each point being one of the plurality of medicinal information, comprising;
a) a data acquisition module configured to obtain prescription data from various sources, including scanned images of prescriptions, product images, QR codes, bar codes, and pharmacy invoices;
b) an optical character recognition (OCR) engine configured to converts scanned documents, images or PDF files into the editable and searchable text, and extracting prescription data, including doctor’s details, prescription date, patient information, drug names, dosages, and duration of therapy;
c) an artificial intelligence module for classifying medicinal drug groups based on the processed prescription data with a predetermined threshold of similarities;
d) a database module configured to map medicine brands and their corresponding prescriptions;
e) a processing module to compare the OCR-predicted data with medicine datasets obtained QR code, barcode, image recognition, and pharmacy invoices; and
f) a linkage module configured to compare and link the predicted prescription data obtained through OCR with medicine datasets obtained from QR codes, barcodes, image recognition, and pharmacy invoices, thereby establishing associations between prescriptions and pharmaceutical medicines.
2. The system as claimed in claim 1, wherein, the data acquisition module includes a smartphone camera for capturing clear images of prescriptions, product packages, and QR codes/barcodes, and further includes a mechanism for detecting details from pharmacy invoices.
3. The system as claimed in claim 1, wherein, the OCR engine employs machine learning techniques to recognize and extract text from scanned images, and the artificial intelligence model utilizes the extracted data to categorize medicinal drug groups with a certain threshold of similarities.
4. The system as claimed in claim 1, wherein, the processing module includes databases for branding information, prescription details, acute and chronic medication categorization, and co-prescribed medications.
5. The system as claimed in claim 1, wherein, the linkage module performs a comparative analysis between the predicted prescription data and medicine datasets obtained from QR codes, barcodes, image recognition, and pharmacy invoices, generating associations and linkages between prescriptions and pharmaceutical medicines.
6. The method for data processing for monitoring pharmaceutical prescriptions at various points in dispensing of prescription, each point being one of the plurality of medicinal information, comprises the steps of:
a) acquiring prescription data from various sources, including scanned images of prescriptions, product image recognition, QR code and barcode detection, and pharmacy invoices;
b) processing the obtained prescription data using an OCR engine to convert it into machine-encoded text;
c) categorizing medicinal drug groups based on the processed prescription data with a predetermined threshold of similarities using an artificial intelligence model;
d) mapping medicine brands that can be co-prescribed using a branding database;
e) comparing the OCR-predicted data with medicine datasets obtained via QR code, barcode, image recognition, and pharmacy invoice; and
f) establishing associations between the prescription data and pharmaceutical medicines based on the comparison results.
7. The method as claimed in claim 1, wherein, the prescription data is acquired by capturing clear images of prescriptions, medicine packaging, QR codes, and barcodes using a smartphone camera and obtaining data from pharmacy invoices through a scanning mechanism.
8. The method as claimed in claim 1, wherein, the OCR engine utilizes machine learning algorithms to recognize and convert handwritten or printed prescriptions into machine-encoded text.
9. The method as claimed in claim 1, wherein, the artificial intelligence model improves the accuracy of OCR predictions for various doctors' prescriptions by using a database of handwritten samples.
10. The method as claimed in claim 1, wherein, associations between predicted prescription data and pharmaceutical medicines are established based on matching information from QR code, barcode, image recognition, and pharmacy invoices.