5.1 Exploring Analytics Output
Last updated
Last updated
The output of the training task will be a detector in which the objects are detected and mentioned in the summary as shown in the following image.
The above image shows the output of the annotation report task. The task details include
Task Name: The name of a task is displayed here.
Created Date: The date of creation of a task is displayed.
Imageset: The name of imageset that is used for executing this task is mentioned here.
No. of Images: Total number of images involved in this process is displayed here.
Detection Type: The type of detection whether it is a "Counting" or a "Defect" task is displayed here.
Completed at: The date on which the task is completed.
Class Labels: The various classes involved in this process are mentioned here.
The user may have a Summary report, Detailed Report, or Share the Task with other team members or any other user. Click on the respective buttons to have the associated report.
The above image shows the summary report of the annotation report task. It gives the detailed information and analysis of the task.
The analysis summary gives the percentage-wise details of the objects detected. The pi-chart will give a graphical representation of objects detected as shown in the following image.
As shown in the above image, the user will have the details of the images analyzed, the objects detected per image, etc. The user may select the class labels and accordingly will get the details of the objects detected. Every image will show the number of objects detected. The user may change the severity level and accordingly see the objects detected.
Manual annotation is possible here. To perform the manual annotation, select the image, click on it to open for manual annotation and perform the annotations on it by adding the class labels as shown in the image below.
After completing the manual annotation, all these objects will be added to the report and the user will get better accuracy of the result.
1) View Report: The user can click on the "View Report" button to see the details of a report. This report can be downloaded in PDF format. This pdf report can be saved on the computer which is used in future if required.
Separate reports will be generated for separate tasks for Defect and Count. The above images show the report generated for the defect type of task.
2) Export Data: The data that is related to the annotation report task can be exported by using the "Export Data" button. ZIP file containing JSON files will be downloaded on the system which the user can open to read.
3) Task Details: The user can see the details of a task by clicking on this button.
The above image shows the output of the detector task which includes the details such as:
Task Name: It displays the name of a task.
Created Date: The date of creation is mentioned here.
Imageset: The name of an imageset used to implement this task is displayed here.
Number of Images: The total number of images : Datepresent in the imageset is displayed here.
Detection Type: The type of detection is displayed.
Completed at: Date of completion of a task is mentioned here.
Status: Status of the task such as "Pending", "Completed" is displayed here.
Detectors: The list of detectors detected through this task is displayed here.
Invite the user to view the task with an email address and click on the "Share" button.
The summary report will give the summary of the task in the following format.
The pie-chart graph will give the details of objects detected in percentage as shown below.
Analysis summary is also available in percentage.
The user is allowed to select the objects as per requirements and accordingly the detailed report will get modified as shown in the image.
The user can also change the severity level to see the number of objects with the same severity level.
The user can download the report by clicking on the "Download" button. The report will be downloaded in pdf and Excel format that can be stored on the computer as well as it can be shared with other team members.
The completion of the download process will be intimated through email.
The user can change the display style of image-wise detailed analysis as shown in the following image.
The user can export the data that can be used for further procedures. Just click on the "Export Data" button to export the data. The imageset will be downloaded on the computer which can be easily opened and saved and shared easily.
The data will be exported or downloaded in the form of ZIP format which is unzipped to view the details.
The files will be in JSON format which contains the details of that particular image. To open the details, open the desired file to see the details. The details will be displayed as below:
