It is a computer system that enables lab technician to upload a microscopic blood smear of a patient to detect if they have ALL and categorize the subtype (pre, pro, early, benign) or if the cell is normal using YOLOv5. Alongside this, The model is used to detect melanoma skin cancer using images of the patient skin. The system has two users:
● The lab technician: Uploads the patient’s blood smear or skin image so the model can do its prediction. Furthermore, they can add patients, and delete patients. Lab technicians can save reports and print reports.
● The admin: Can update, delete, and add lab technicians
● Sign up
● Login
● Add new patient
● Add new lab technician
● Show report by patient ID
● Edit patient
● Delete patient
● Delete doctor
● Show patient details
● Upload image
● Log out
● Print report
● Save report
- Constraints:
● Doctor username must be unique
● Password and confirm password fields must be identical
● Password must be mix of letters and number and 6 characters long
● Gender is in dropdown list to ensure consistency
● Mobile number must begin with 011, 012, 010, or 015
● Mobile number must be 11 numbers
● Ensure that the username and password are correct before logging in
● The system detects ALL subtypes accurately and fast using an object detection model called YOLOv5 which is the main goal of the project.
● A generated report of the patient’s case will be generated and it can be printed so the patient can take it to their physicians.
● The patient can have more than one image on the database.
● Each lab technician has their patients and can’t access the patients of other technicians.
● The system shows the predicted class of the ALL and shows the bounding boxes with the confidence level of the prediction.
This study evaluates object detection model to be specific on ALL image dataset. The bone marrow laboratory at Taleqani Hospital provided the images for this dataset (Tehran, Iran). This dataset included 3256 PBS pictures from 89 individuals who were thought to be ALL and whose blood samples were properly processed and stained by professional laboratory personnel. This dataset is separated into the benign and malignant classes. Hematogenes make up the first group, whereas the ALL group, which includes the three malignant lymphoblast subtypes Early Pre-B, Pre-B, and Pro-B ALL, makes up the latter. All of the photos were taken with a Zeiss camera at a 100x magnification in a microscope, and they were all saved as JPG files. The specific types and subtypes of these cells were identified by a professional using the flow cytometry instrument. Segmented photos following colour thresholding-based segmentation in the HSV colour space are offered. However, segmentation made a lot of details to be lost leading to inaccuracies, therefore, the original images were used . ((((((((add images)) pg 25
- Melanoma Skin Cancer Dataset
- Dataset from Kaggle 10000 pictures are included in the Melanoma Skin Cancer Dataset.. 9600 images make up the dataset, whereas 1000 images are used to evaluate the model. It is subdivided into two classes: malignant and bening. The dataset is found in Kaggle with no bounding boxes. Figure 2 displays samples of data from Kaggle. (((images )))
- Dataset from Roboflow The same exact dataset is found in Roboflow but only 4300 pictures are found with the bounding boxes automatically assigned to the objects. Figure 3 illustrates samples of the same Melanoma Skin Cancer dataset but with bounding boxes from Roboflow ( images)
- Functional requirements
-
Login Actor: User (Doctor) Precondition: Open application and have active internet connection. 38 Description: The Doctor will start entering his information such as username and password. Postcondition: Access his personal account & other services.
-
Add Patient Actor: User (Doctor) Precondition: The Doctor must be logged in. Description: The doctor should enter the patient name, gender, phone number, date of birth of the patient to be created. Postcondition: New patient information will be displayed in the table and saved in the firebase.
-
Update Patient Actor: User (Doctor) Precondition: The Doctor must be logged in and know the name or id of the user that he wants to update. Description: The doctor should enter patient id and the updated patient fields of the patient to be updated. 39 Postcondition: Patient information will be updated in the firebase and displayed in the table on screen.
-
Delete Patient Actor: User (Doctor) Precondition: The Doctor must be logged in and know the name or id of the user that he wants to delete. Description: The doctor should enter the id of the patient to be deleted. Postcondition: Patient information will be deleted in the firebase and removed from the table on screen.
-
Search Patient Actor: User (Doctor) Precondition: The Doctor must be logged in and patient data exists. Description: The doctor should enter the id of the patient and search patient data. Postcondition: Patient information will be displayed in a table.
-
Browse Image Actor: User (Doctor) Precondition: The Doctor must be logged in. Description: The doctor should enter the patient id and patient name and browse blood smear image from pc. Postcondition: The system detects the cancer type and location of cancer cells in a blood smear image.
-
Generate Report Actor: User (Doctor) Precondition: The Doctor must be logged in and a blood smear image is uploaded. Description: When the doctor presses the “Generate Report” button, the system redirects him to the Examination Report window to view the generated report. Postcondition: Detailed report of the patient’s blood smear will be displayed in the Examination Report window.
-
Save Report Actor: User (Doctor) Precondition: The Doctor must be logged in and a report is generated. Description: When the doctor presses the “Save Report” button, the report is saved in the firebase. Postcondition: The generated report and patient information are saved in the firebase.
-
View report numbers Actor: User (Doctor) Precondition: The Doctor must be logged in and patient exits. Description: The doctor is redirected to show a report window and all the reports numbers of the patient will be displayed in the table. Postcondition: The report numbers of reports of the patient will be displayed in the table.
-
View report Actor: User (Doctor) Precondition: The Doctor must be logged in and patient exits. Description: The doctor is redirected to the Examination Report window and the patient report is displayed on screen. Postcondition: The patient report will be displayed on screen.
12.Save report as pdf Actor: User (Doctor) Precondition: The Doctor must be logged in and the patient and report number exists. Description: The doctor will choose the folder to which the report file will be saved as pdf on pc and the report file will be saved. Postcondition: Report is saved as pdf and doctor views report file of patient as pdf file.
13.Logout Actor: User (Doctor) Precondition: The Doctor must be logged in. Description: When the doctor presses the “Logout” button, he will be directed to the login window. Postcondition: Doctor will log out successfully & won’t use the app without login again.
14.Add Doctor Actor: User (Admin) Precondition: Must be logged in. Description: The admin should enter new doctor id, password and confirm password then create doctor. Postcondition: New doctor information is saved in the firebase.
15.Delete Doctor Actor: User (Admin) Precondition: Must be logged in, doctor exists. Description: The doctor should enter id, password and confirm password of the doctor to be deleted. Postcondition: Doctor information will be deleted in the firebase.
16.Update Doctor Actor: User (Doctor) Precondition: Must be logged in and the doctor exists. Description: The doctor should enter his old password, new password and confirm password to update his information. Postcondition: Doctor information will be updated in the firebase and the system outputs “Password updated” message on screen.
- Non-functional requirements ● Performance: The response time for the system is few milliseconds depending on the device it is working on.
● Maintainability: It is easy to maintain and update the system anytime because of the high system integrity.
● Portability and compatibility: The system can work on any RAM capacity equals or higher that 8 gigabyte. Moreover, it can be opened using any type of pc and any type of operating system.
● Usability: Due to the user-friendly design of the system and the utilization of the graphical user interface it is extremely easy for the users to use our system.
● Capacity: unlimited users can concurrently use it.
● Security: Only authorized users can access the system, and only authorized users for a specific patient can access the patient information.
● The user shall be able to login to the system
● The user shall be able to add, delete, view and update patient information.
● The user shall be able to browse a blood smear image and examine the image for skin or leukemia cancer detection.
● The user shall be able to generate a report of the biopsy image.
● The user shall be able to save the generated report as pdf on pc.
● The user shall be able to view the reports numbers of the patient and choose which report to view.
● The user shall be able to update his password.
● The user shall be able to logout of the system.
● PC: The application works on pc that runs any operating system such as windows, linux ,mac os.
● Internet connection: This app uses a firebase to store data so you need to be connected to the internet to fetch, view and modify it.
● The role between the responsible lab technician for a specific patient and other different types of users must be defined.
● RAM must be at least 8 gigabyte for the system to work.
- Add Doctor with incorrect password OR password and confirm password are not identical
This allows the admin to add and delete doctors by providing the username, password and confirm password of the doctor. The password should be less than 6 characters and contain letters and numbers. If the admin doesn’t follow these constraints, the system outputs the message “Password should be at least character long and contain letters and numbers” on the screen. The fields of Password and Confirm Password must be identical
- Add Doctor Successfully
This allows the admin to add and delete doctor by providing the username, password and confirm password of the doctor. If the admin provides username, password and confirms password following password constraints, doctor is added successfully in the firebase and system outputs “Doctor is added Successfully” message on screen.
-
Add Doctor with incomplete data
-
Add Doctor that already exists This allows the admin to add and delete doctor by providing the username, password and confirm password of the doctor. If the admin provides the username of the doctor that already exists when adding a new doctor, the system outputs the “User already exists” message on screen.
- Update Doctor with incorrect old password This screen allows the doctor to update his password by providing his old password, new password and confirm password. If the Doctor provides an incorrect old password, the system outputs a “No such doctor” message on screen.
- Update Doctor with incorrect password This screen allows the doctor to update his password by providing his old password, new password and confirm password. The password should be less than 6 characters and contain letters and numbers. If the doctor doesn’t follow these constraints, the system outputs the message “Password should be at least character long and contain letters and numbers” on the screen.
- Update Doctor with unmatched new and confirm password
This screen allows the doctor to update his password by providing his old password, new password and confirm password. If the new and confirm password are unmatched, System outputs “New password and confirm password are not matched” message on screen.
- Update doctor with incomplete data
This screen allows the doctor to update his password by providing his old password, new password and confirm password.
If the doctor provides incomplete data, the system outputs a “Please complete the data” message on screen.
- Update Doctor successfully
This screen allows the doctor to update his password by providing old password, new password and confirm password. If the doctor provides the old, new and confirm password following password constraints, the doctor's password is updated successfully in the firebase and the system outputs “Password updated” message on screen.
- Add Patient
This screen allows doctor to add by providing name, gender, phone number and date of birth of patient, update by providing id and the new and old data of patient, delete by providing id of patient, search by providing id of patient, show reports numbers of patient, show data of all patients in the table and logout.
If the doctor provides name, gender, phone number and Date of birth following constraints and presses the create patient button, a new patient is created successfully in the firebase and the system displays new patient data on the table.
- Add patient with incorrect date of birth
This screen allows doctor to add by providing name, gender, phone number and date of birth of patient, update by providing id and the new and old data of patient, delete by providing id of patient, search by providing id of patient, show reports numbers of patient, show data of all patients in the table and logout.
The date of birth should be in dd/mm/yyyy format. If the doctor doesn’t follow these constraints, the system outputs the message “Enter date in dd/mm/yyyy” format on the screen.
- Add Patient with incorrect phone number
This screen allows doctor to add by providing name, gender, phone number and date of birth of patient, update by providing id and the new and old data of patient, delete by providing id of patient, search by providing id of patient, show reports numbers of patient, show data of all patients in the table and logout.
If the doctor doesn’t follow phone number constraints, the system outputs the message “Invalid phone number” on the screen.
- Incomplete data If data is incompleted show error message “Please complete the data”
- Update Patient
This screen allows doctor to add by providing name, gender, phone number and date of birth of patient, update by providing id and the new and old data of patient, delete by providing id of patient, search by providing id of patient, show reports numbers of patient, show data of all patients in table and logout. Doctor provides the data to be modified in patient and id of patient to be updated, patient data will be updated from firebase and the row containing patient data will be updated with the modified patient data in the table.
- Choose Section This screen allows the doctor to choose whether to go to the patient management section of the system to create new patients and access data and report of existing patients or to go to the sample examination section to examine biopsy images.
- Choose Cancer Section This screen allows doctors to choose whether to go to the leukemia cancer detection section or to the skin cancer detection section.
- peripheral blood smear examination
This screen allows doctor to perform examination on blood smear image by providing patient id and name and upload image of blood smear from device by pressing button browse then the system will perform detection on image and identify the location of cancer cells. When the doctor presses the generate report button, the system generates a report of the case.
- Examination Report of Skin Cancer This screen allows the doctor to view report of the examined skin biopsy, save report to firebase along with patient data and save data as pdf in the desired location in pc.
- Examination Report of Leukemia This screen allows the doctor to view the report of the examined peripheral blood smear, save the report to firebase along with patient data and save data as pdf in the desired location on the PC.
- The Generated Report as pdf The screen shows the generated report of blood smear image as pdf on pc.
- Skin biopsy examination
This screen allows the doctor to perform an examination on skin biopsy image by providing patient id and name and upload image of blood smear from device by pressing browse button then the system will perform detection on image and identify the location of cancer cells. When the doctor presses the generate report button, the system generates a report of the case.
- Login
This screen allows the doctor and admin to login to the system by providing an id and password. The user must enter the correct id and password to access the system otherwise a message will be shown saying “Wrong doctor password or ID
- View saved report This screen allows the doctor to view the reports numbers of the saved reports of the patient and to view and save the report as pdf by providing patient id and report number and press the show report button. After pressing the show report button , the doctor will be redirected to the Examination Report window and the report will be displayed on screen.
https://drive.google.com/file/d/1OWJ9XI6PzgfKCVEKYYGiJ4wlAkK6yql0/view?usp=sharing