OCR document self-learning
1 Introduction
OCR document self-study , seeing this title, many people's first reaction is, I don't know how to do algorithms?
In fact, if you carefully read the introduction of OCR document self-study, you will know that this product is mainly for enterprises and individual developers who have "no algorithm foundation".
So, seeing this, don't you feel relieved?
Now that there is no algorithm-based blessing, then we can safely and boldly enter today's product evaluation environment.
I have always been adhering to the attitude of using Alibaba Cloud products with a learning attitude, even if it is an evaluation product, it will not change my mentality.
Of course, I will also "destroy" this product from the perspective of the user. After all, rather than writing out the experience, I believe that the students of the OCR team hope to bring a different experience... Hehe...
2. Unboxing experience
First of all, what catches the eye is the overview page of OCR document self-study
On the overview page, we are familiar with:
quick start guide,
model selection list,
service activation and permissions,
exclusive customer service DingTalk group.
Of course, since there is a fee, there is also the billing details ,
click the billing details button, and jump to the payment description document, which will let you spend your money clearly.
3. My task menu
3.1 Create a task
Switch to the My Task menu, and similarly, the list will display the data information of the task, as well as various template information, as follows
Here, we create a task, that is, click the create task button to enter the create task page
On the task creation page, the blank space on the right side is so large, I don’t know if you can get used to it, anyway, it’s my habit.
Next, we select a custom form template, fill in the information, and click the Create button
Sure enough, the result didn't disappoint me, it reminded me that the service was not activated
In this case, you must click the button to activate the custom form template service
Click the button to activate immediately, and see this page, which means that the activation is successful.
Go back to our task creation page and click the Create button directly to create a task successfully.
3.2 Editing tasks
My task list, click the edit button to enter the edit task page
Task administrator: If you only bind one user, then click X, then this is not a delete button, but a drop-down list, as shown below:
Advanced options: Expand the advanced options, click on the blank space, and the advanced options will be automatically collapsed... This function is very nice???
Before clicking
after click
3.3 Create a template
After creating the task, we will then create the template.
1. Upload template
When I click to replace the template sample image, the image is directly deleted and restored to the original state...
If this is a wrong operation, it will be really cool.
I am re-uploading one, this time I changed to drag and drop,
Second, select the reference field
Click the next button, the result, still did not disappoint me, prompting missing COR information parameters...
I just want to ask weakly, for user Xiaobai,
- What is OCR User Parameters,
- What are the OCR user parameters,
- What does the prism_wordsinfo parameter represent? ?
If you know, welcome to leave a message in the comment area.
When uploading an unsupported format, in the error prompt pop-up window,
what is the difference between the actual function meaning of the know button and the close button? ?
A pop-up window, three "close" buttons (X, got it, close), is this for aesthetics? ?
After unremitting efforts, I uploaded the PDF file, and I can proceed to the next step normally, namely: select the reference field tab
On this page, similarly, the left menu bar is automatically folded, and this function is fine.
However, the PDF document I uploaded has 11 pages, and on this page, only the content of the first page is recognized. Is this set on purpose?
Then, I dragged the document to an appropriate location, clicked Save, and it was successfully saved normally.
3. Configure the identification field
In the configuration identification field tab, I clicked the comparison button, and the feeling of ghosting in this document is awesome.
Then I added some information, and after clicking Next, I can perfectly enter the template test and release tab
4. Template testing and release
Let's continue to upload pictures, and then click the publish template and start API call button to see if the release is successful.
Seeing the information here means that our template has been published successfully.
Let's click back to details again to see the information of the template
No problem, this is the template we just created.
4. Use the menu online
Go to the online use menu, select the template that needs to be published, there is nothing wrong with this.
Let's continue to upload the document and see the effect
In the process of uploading the document, did you find that the title of the document is quite small, and it also exceeds the outside of the frame.
This seems to need modification.
Click on API integration to jump to the API integration documentation, good guy, I thought this API integration was a function, but I didn’t expect it to be a document .
5. Data monitoring menu
Switch to the data monitoring menu, according to the task, you can query the information of this task
There is no problem with this function.
However, there is a problem here. If my "task" is not selected, and I click the query button directly, a prompt message will be given, as shown in the figure below. Since
"task" is a mandatory option, can I first give the identification of the required item? How to prevent everyone from making a mistake once?
5. Toolbox
Field Type Management → My Fields: Click the Add Field Type button to fill in the information and add
Field Type Management → My Fields: List Display
Similarly, the content of the dictionary can be downloaded on the viewing page
6. Summary
In the overall use process, it feels quite convenient to me.
And as a novice user with no algorithm foundation, except for a few prompts that I don’t understand, other operating procedures and template types are very simple and rich.
In fact, as a product, in addition to rich functions, it also needs to be easy to operate, allowing new users to get started directly,
saving users precious time by eliminating cumbersome steps.
Although in the AI era, algorithms need to be mastered by technicians,
OCR document self-study is positioned as a basic template without algorithms for enterprises and users.
Then, some copywriting tips and the use of functions must not propose too many technical related After all,
the user of the product is not necessarily a technical er, but more likely to be a Boss, or a novice user who does not understand technology.
I am a small fish :
- CSDN blog expert ;
- Aliyun expert blogger ;
- 51CTO blog expert ;
- 51 certified lecturer, etc .;
- Certified gold interviewer ;
- Job interview and training planner ;
- Certified expert bloggers in several domestic mainstream technical communities ;
- Winners of the first and second prizes in the evaluation of various mainstream products (Alibaba Cloud, etc.) ;
I am Xiaoyu, follow me, and I will show you more professional and prefaced test development techniques.