Document Digitization and Medical Record Retrieval in Healthcare
Project Summary Patients are challenged in their daily lives by maintaining their multiple medical records in huge number of piles. The resources available for them are limited and localized. Each record is filled with information about the patient such as their identity, med
2025-06-28 16:32:11 - Adil Khan
Document Digitization and Medical Record Retrieval in Healthcare
Project Area of Specialization Artificial IntelligenceProject SummaryProject Summary
Patients are challenged in their daily lives by maintaining their multiple medical records in huge number of piles. The resources available for them are limited and localized. Each record is filled with information about the patient such as their identity, medical history, and laboratory results. Given the importance of health records, it is essential to keep them up to date so that optimal care can be provided when necessary. However the vast amount of records makes updating them a difficult job. Doctors face difficulty to check the medical history for their patients. Hospitals often face challenges in sifting through voluminous healthcare records and extracting relevant information from physical documents.
Fortunately there is a system needed for health records digitization, which can streamline information and keep it current. A system that merges the medical history of the respective patients into one single document and update the medical history of the patient automatically to cloud-based storage.
Document Digitization and Medical Record Retrieval in Healthcare is a compatible with all OPERATING SYTEM application with a goal to automate the collection of medical data through medical reports. The data will be processed and retrieved in a format by OCR after the uploading of the medical report, analysis of extracted medical data through OCR medical history of the patient will be updated on the basis of matching algorithm. The output will be easily transformed into human readable documents by means of standard techniques allowing the users to graphically compare the results against the input image of the document.
Project ObjectivesProject Objective
We are going to implement a mobile application in which user will upload a scanned document of his/her history in the mobile application and whenever needed it is exported to the respective Doctor, irrespective of the hospital. Now, the reports can be in hardcopy or the patients will be accessing his or her report through a URL provided by the hospital. The reports’ PDF will be uploaded if it’s present in hardcopy or if URL is provided the image of URL will be uploaded and medical data will then be available through web scraping. Now here our first objective is to design an OCR engine that has the capability to gather all the data written in the report. This would be a challenging task as the reports are printed through different printers like inkjet, laser and dot matrix printers, and to recognize the characters in that format will require some algorithm. Our second challenge would be to design a module for placement of medical data in the right domain and this will require some matching algorithm. Here the challenging task is that reports can be different; there is no universal format for medical report. All the hospitals produce medical report in their own format. So, our matching algorithm must be that much efficient that it places the extracted and analyzed medical data in right places, like it should have the ability to recognize test name, result and unit.
Project Implementation MethodMETHODOLOGY:
The main stages that allow the processing of a printed page containing the laboratory results are: Image preprocessing, in which the document readability is enhanced; layout analysis, in which the document layout is analyzed in order to identify columns and rows containing the information to be extracted, data extraction and classification, in which text returned by the OCR is analyzed syntactically and semantically exportation in a format of the extracted data, and finally updating of the medical history.
I) Preprocessing: In this phase the image is prepared for subsequent processing steps. In particular, equalization, binarization and suppression of long lines are required to ease layout analysis and OCR.
II) Layout analysis: In the layout analysis phase, the image is subdivided into blocks; text rows are processed by the OCR, and the text is compared with a list of column headers. The column headers identify those table columns that contain pertinent medical data.
III) Data extraction: In this phase data are extracted from table cells. The cells are processed by the OCR, which is configured to recognize a different set of characters according to the cell type.
IV) Exportation in a format and Automated Updation in medical record: In this phase, the data extracted from the document are saved in an output file. This output will conceive in order to share medical laboratory results in a simple manner. Other than integration into databases, the output can be easily transformed into human readable documents by means of standard techniques. It should be noted, however, that the information on the coordinates of the extracted information is preserved in the results, so allowing the users to graphically compare them against the input image of the document.
Web Scrapping Methodology: When the patient will upload a receipt of his test, including URL, ID and password to access the test reports, the OCR will retrieve this information and use web scrapping to obtain the test results the methodology of web scrapping will be as follows:
(i) Find the URL that you want to scrape: First, you should understand the requirement of data according to your project. A webpage or website contains a large amount of information. That's why scrap only relevant information. In simple words, the developer should be familiar with the data requirement.
(ii) Inspecting the Page: The data is extracted in raw HTML format, which must be carefully parsed and reduce the noise from the raw data. In some cases, data can be simple as name and address or as complex as high dimensional weather and stock market data.
(iii) Write the code: Write a code to extract the information, provide relevant information, and run the code.
(iv) Store the data in the file: Store that information in database. This stored date will update the record of patient’s medical history.
Benefits of the Project- Smart Product.
Patients will be able to upload their past medical reports through OCR scanner.
- Artificial Intelligence.
Based on the medical history and the new conditions, there will be trends and predictions for the patient’s health making it easier for doctors to understand patient’s health.
- General Prescriptions.
The application will be able to propose general prescriptions, dosage, and time duration for its effectiveness based on the symptoms entered by the patient.
- Secure Database.
Patients’ records will be secured and will not be shared until and unless patient himself/herself wants to share it with someone.
- Sharing with Doctors.
If the patients want to share the record with concerned doctor, the patient will generate a request with the credentials or pass key.
Technical Details of Final DeliverableTechnical Details of Final Deliverable
AUTHENTICATION:
It will ask for the user’s authenticity to get services from the system module by giving its own authentication keys
while sign up.
SYSTEM MODULE:
It will be the main home page of the application from which user can upload or view his/her medical record
according to his needs. Also he/she can export its data to make sure that his new doctor knows about his history.
UPLOAD RECEIPT OF MEDICAL TESTS:
Whenever we take tests, hospital gave us the receipt on which the id, password and URL is printed on which the
report has to be uploaded so it will take picture or pdf of receipt.
UPLOAD REPORTS OF MEDICAL TESTS:
Patients also have their record as PDF file or as a hard copy so here patient can upload his/her files.
OCR:
Both the above uploading module will go to OCR and OCR will extract the relevant information and use it
according to the system needs.
DATA PROCESSING:
Data will now be extracted and being analyzed and categorized according to the specific requirements.
WEB SCRAPING:
When the system receives the receipt information then by using web scraping, it will extract and store the content
in the system.
UPDATE RECORD:
The records will be automatically updated after every new record is added to the Application.
| Item Name | Type | No. of Units | Per Unit Cost (in Rs) | Total (in Rs) |
|---|---|---|---|---|
| Total in (Rs) | 70000 | |||
| Sheet feed scanner | Equipment | 1 | 15000 | 15000 |
| Workstation | Equipment | 1 | 40000 | 40000 |
| Cloud Vision | Equipment | 1 | 8000 | 8000 |
| Hospitals/laboratory Visits | Miscellaneous | 1 | 2000 | 2000 |
| USB/Cables | Miscellaneous | 1 | 2000 | 2000 |
| Printing and stationary cost | Miscellaneous | 1 | 3000 | 3000 |