Smart Use Cases & Code Samples

PixLab Console

Version 1.9.72 (Release Notes)

Smart Use Cases & Code Samples

The code samples set are of practical usage, production ready snippets available in various programming languages intended to familiarize the reader with the PixLab API. For an introduction course see Getting Started with PixLab, The PixLab Rest API and The Endpoints Reference Guide. You’re welcome to copy/paste and run these examples to see the API in action.

These examples and dozen of others are available on the PixLab Github Sample Repository.

Scan US Driver's License

In this sample, we'll learn how to scan a given US Driver's License issued by any state despite the huge difference in shape & form which varies from one US state to another. This is done using a the DOCSCAN API Endpoint ↗ which also extracts (crop) any detected face within the target document with the mandatory fields that is the driver's name, address & license number, and many others. A typical input US Driver's License specimen, and the extracted output fields should look like the following after processing:

Input US Driver's License

US Driver license Specimen

Extracted License Fields

US Driver License Scan Output

DOCSCAN is the sole API endpoint needed for such a task. It extracts the mandatory fields including driver's name, address & license number regardless of the issuing US state, and does face extraction automatically for you.
PixLab recommend that you connect your AWS S3 bucket via the console so that, any extracted face or MRZ crop is automatically stored on your S3 bucket rather than the PixLab one. This feature should give you full control over your analyzed media files.
Refer to the DOCSCAN API documentation for additional information such as the set of scanned fields, where face crops are stored, how to process PDF documents instead of images and so forth.

To learn more about the DOCSCAN API endpoint capabilities, including the list of supported documents & code samples, please refer to the following blog posts:

The Code

This code and dozen of others are available on the PixLab Github Repository.

Python
PHP

Scan a government issued Passport

In this production-ready code samples, we'll scan any officially issued ID document such as: Passport, Visa, Residence Card, Driver's License, or any type of ID card with or without MRZ from various countries ↗ using a single API endpoint. With support for over 11,000 types of ID documents from 197+ countries, the DOCSCAN API Endpoint ↗ not only provides highly accurate text scanning but also extracts (crops) any detected face within the target document. A typical input user uploaded Passport first image, and the extracted output fields should look like the following after processing:

Input Passport Uploaded Image Specimen

Passport Specimen

Parsed MRZ Fields

Docscan Passport Scan Output

Input US Driver's License

US Driver license Specimen

Extracted License Fields

US Driver License Scan Output

DOCSCAN is the sole endpoint needed for such a task. It support various ID cards besides Passports, US Driving Licenses & Visas, and does face extraction automatically for you.
PixLab recommend that you connect your AWS S3 bucket via the dashboard so that, any extracted face or MRZ crop is automatically stored on your S3 bucket rather than the PixLab one. This feature should give you full control over your analyzed media files.
Refer to the DOCSCAN API documentation for additional information such as the set of scanned fields, where face crops are stored, how to process PDF documents instead of images and so forth.

To learn more about the DOCSCAN API endpoint capabilities, including the list of supported documents & code samples, please refer to the following blog posts:

The Code

This code and dozen of others are available on the PixLab Github Repository.

Python
PHP

Mimic Snapachat Filters

In this sample, we shall mimic the famous Snapchat filters on a set of human faces. Actually, producing those filters is quite easy and require only three API endpoints: facelandmarks, smartresize & merge plus a bunch of other optional endpoints for cool background effects if desired.

1957 Elia Kazan film “A Face in the Crowd.” Photo courtesy of Warner Bros. Dog members

Dog members

FACELANDMARKS is the analysis endpoint that must be called first. It tries to detect all present human faces, extract their coordinates and more importantly, it output the landmarks for each face. That is, the coordinates for each facial member such as the eye, bone, nose, mouth, chin, lip, etc. We'll use these coordinates later to composite stuff such as the famous flower crown, dog members, etc. on top of the desired facial region. Find out more information about facelandmarks here.

SMARTRESIZE is called in order to fit the image to be composited (i.e. The flower crown) to the desired dimension such as the bone width.

MERGE is the processing endpoint we call next. It expect a series of coordinates (X, Y) and a list of images to be composited on top of the target image. More about merge here.

You may have noticed that PixLab ships with dozen of filters such as grayscale, noise, blur, chop, despeckle, deskew, edge, emboss, rotate, flip, etc. Don't forget to use one of them for some cool background effects if desired.

The Code

This code and dozen of others are available on the PixLab Github Repository.

Python
PHP

Detect all human faces present in a given image and extract each one of them

In this sample, we'll perform human face detection using the FACEDETECT endpoint, and finally extract each detected face via CROP.

input faces

Input Picture

face crop

After Crop

To perform such operation, two endpoints are actually needed: One for the analysis phase, the other for processing:

FACEDETECT is the analysis endpoint that must be called first. It outputs the rectangle coordinates for each detected face in a given image or video frame. This endpoint, its expected parameter, and JSON output result is documented here.
CROP is called later after we obtain the rectangle coordinates for all the detected faces. In which case, we simply pass these coordinates to crop untouched and the target region of the image where each face is located is extracted. Find out more information about the CROP endpoint here.
Alternatively, if you prefer marking the face(s) instead of extracting it, you can rely on the DRAWRECTANGLE endpoint to perform such task.

The Code

This code and dozen of others are available on the PixLab Github Repository.

Python
PHP

Dynamically create a new image & draw some text on top

In this example, we'll dynamically create a 300x300 pixels, PNG image with a yellow background and draw some text on top. This is very useful, if you want for example to challenge bots on your platform by implementing a minimalistic yet powerful Captcha-like challenge. A typical generated image should look like the following after processing:

Draw after processing

Dynamically created using NEWIMAGE, text drawn via DRAWTEXT

A blank image is requested first via call to NEWIMAGE which export to PNG by default but you can change the output format at request. We set the image height, width and the background color respectively to 300x300 with a yellow background color.
Note, if one of the height or width parameter is missing (but not both), then the available length is applied to the missing side and if you want a transparent image set the color parameter to none.
We finally DRAW our text at the center of the newly created image using the WOLF Font, black color and 35 px font size. One could also draw lines, a rectangle, and other shapes for example to surround faces, merge with other images and so on...

The Code

This code and dozen of others are available on the PixLab Github Repository.

Python
PHP