As you can see, if your application accepts user-generated content, moderation might be a requirement that you can\u2019t ignore. There are different ways moderation can be carried out:<\/p>\n
- \n
- Individual driven<\/strong>\u00a0– an example of this is a website that has admins that moderate the content. The website might work by either restricting the display of any uploaded content until it has been approved by an admin or it might allow immediate display of uploaded content, but have admins who constantly check posted content. This method tends to be very accurate in identifying inappropriate content, as the admins will most likely be clear as to what is appropriate\/inappropriate for the medium. The obvious problem with this is the human labor needed. Hiring moderators might get costly especially as the application\u2019s usage grows. Relying on human moderators can also affect the app\u2019s user experience. The human response will always be slower than an automated one. Even if you have people working on moderation at all times, there will still be a delay in identifying and removing problematic content. By the time it is removed, a lot of users could have seen it. On systems that restrict showing uploaded content until it has been approved by an admin, this delay can become annoying to users.<\/li>\n
- Community driven<\/strong>\u00a0– with this type of moderation, the owner of the application puts in place features that enable the app\u2019s users to report any inappropriate content e.g. flagging the content. After a user flags a post, an admin will then be notified. This also suffers from a delay in identifying inappropriate content from both the community (who might not act immediately the content is posted) and the administrators (who might be slow to respond to flagged content). Leaving moderation up to the community might also result in reported false positives as content that is safe is seen by some users as inappropriate. With a large community, you will always have differing opinions, and because many people will probably not have read the Terms and Conditions of the medium, they will not have clear-cut rules of what is and isn\u2019t okay.<\/li>\n
- Automated<\/strong>\u00a0– with this, a computer system usually using some machine learning algorithm is used to classify and identify problematic content. It can then act by removing the content or flagging it and notifying an admin. With this, there is a decreased need for human labor, but the downside is that it might be less accurate than a human moderator.<\/li>\n
- A mix of some or all the above methods<\/strong>\u00a0– Each of the methods described above comes with a shortcoming. The best outcome might be achieved by combining some or all of them e.g. you might have in place an automated system that flags suspicious content while at the same time enabling the community to also flag content. An admin can then come in to determine what to do with the content.<\/li>\n<\/ul>\n
<\/a>A Look at the Imagga NSFW Categorization API<\/h2>\n
Imagga makes available the\u00a0NSFW (not safe for work<\/a>) Categorization API<\/strong>\u00a0that you can use to build a system that can detect adult content. The API works by categorizing images into three categories:<\/p>\n
- \n
- nsfw<\/strong>\u00a0– these are images considered not safe. Chances are high that they contain ponographic content and\/or display nude bodies or inappropriate body parts.<\/li>\n
- underwear<\/strong>\u00a0– this categorizes medium safe images. These might be images displaying lingerie, underwear, swimwear, e.t.c.<\/li>\n
- safe<\/strong>\u00a0– these are completely safe images with no nudity.<\/li>\n<\/ul>\n
The API works by giving a confidence level of a submitted image. The confidence is a percentage that indicates the probability of an image belonging to a certain category.<\/p>\n
To see the NSFW API in action, we\u2019ll create two simple programs that will process some images using the API. The first program will demonstrate how to categorize a single image while the second will batch process several images.<\/p>\n
<\/a>Setting up the Environment<\/h2>\n
Before writing any code, we\u2019ll first set up a\u00a0virtual environment<\/a>. This isn\u2019t necessary but is recommended as it prevents package clutter and version conflicts in your system\u2019s global Python interpreter.<\/p>\n
First, create a directory where you\u2019ll put your code files.<\/p>\n
$ mkdir nsfw_test<\/pre><\/p>\n
Then navigate to that directory with your Terminal application.<\/p>\n
$ cd nsfw_test<\/pre><\/p>\n
Create the virtual environment by running:<\/p>\n
$ python3 -m venv venv<\/pre><\/p>\n
We\u2019ll use Python 3 in our code. In the above, we create a virtual environment with Python 3. With this, the default Python version inside the virtual environment will be version 3.<\/p>\n
Activate the environment with (on MacOS and Linux):<\/p>\n
$ source venv\/bin\/activate<\/pre><\/p>\n
On Windows:<\/p>\n
$ venv\\Scripts\\activate<\/pre><\/p>\n
<\/a>Categorizing Images<\/h2>\n
To classify an image with the NSFW API, you can either send a GET request with the image URL to the
\/categories\/<categorizer_id><\/code> endpoint or you can upload the image to\/uploads<\/code>, get back aupload_id<\/code> value which you will then use in the call to the\/categories\/<categorizer_id>\/<\/code> endpoint. We\u2019ll create two applications that demonstrate these two scenarios.<\/p>\nProcessing a Single Image<\/h3>\n
The first app we\u2019ll create is a simple web application that can be used to check if an image is safe or not. We\u2019ll create the app with\u00a0Flask<\/a>.<\/p>\n
To start off, install the following dependencies.<\/p>\n
$ pip install flask flask-bootstrap requests<\/pre><\/p>\n
Then create a folder named\u00a0
templates<\/code>\u00a0and inside that folder, create a file named\u00a0index.html<\/code>\u00a0and add the following code to it.<\/p>\n{% extends \"bootstrap\/base.html\" %}\r\n\r\n{% block title %}Imagga NSFW API Test{% endblock %}\r\n\r\n{% block navbar %}\r\n\r\n<nav class=\"navbar navbar-inverse\" role=\"navigation\">\r\n<div class=\"container\"><a class=\"navbar-brand\" href=\"{{ url_for('index') }}\">NSFW API Test<\/a><\/div>\r\n<\/nav>{% endblock %}\r\n\r\n{% block content %}\r\n<div class=\"container\">\r\n <div class=\"row\">\r\n <div class=\"col-md-8\">\r\n <form action=\"\" method=\"POST\">\r\n\r\n <div class=\"form-group\">\r\n <label for=\"image_url\">Image URL<\/label>\r\n <input type=\"url\" id=\"image_url\" name=\"image_url\"\/>\r\n <button class=\"btn btn-primary\" type=\"submit\">Submit<\/button>\r\n <\/div>\r\n\r\n <\/form>\r\n <\/div>\r\n <\/div>\r\n {% if image_url %}\r\n <div class=\"row\">\r\n <div class=\"col-md-4\"><img class=\"img-thumbnail\" src=\"{{ image_url }}\" \/><\/div>\r\n <div class=\"col-md-4\">{{ res }}<\/div>\r\n <\/div>\r\n {% endif %}\r\n\r\n<\/div>\r\n{% endblock %}\r\n<\/pre>\nIn the above code, we create an HTML template containing a form that the user can use to submit an image URL to the Imagga API. When the response comes back from the server, it will be shown next to the processed image.<\/p>\n
Next, create a file named\u00a0
app.py<\/code>\u00a0in the root directory of your project and add the following code to it. Be sure to replace\u00a0INSERT_API_KEY<\/code>\u00a0and\u00a0INSERT_API_SECRET<\/code>\u00a0with your Imagga API Key and Secret. You can\u00a0signup<\/a>\u00a0for a free account to get these credentials. After creating an account, you\u2019ll find these values on your\u00a0dashboard<\/a>:<\/p>\nfrom flask import Flask, render_template, request\r\nfrom flask_bootstrap import Bootstrap\r\nimport requests\r\nfrom requests.auth import HTTPBasicAuth\r\n\r\n\r\napp = Flask(__name__)\r\nBootstrap(app)\r\n\r\n# API Credentials. Set your API Key and Secret here\r\nAPI_KEY = 'acc_xxxxxxxxxxxxxxx'\r\nAPI_SECRET = 'xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx'\r\nENDPOINT = 'https:\/\/api.imagga.com\/v2\/categories\/nsfw_beta'\r\nauth = HTTPBasicAuth(API_KEY, API_SECRET)\r\n\r\n\r\n@app.route('\/', methods=['GET', 'POST'])\r\ndef index():\r\n image_url = None\r\n res = None\r\n\r\n if request.method == 'POST' and 'image_url' in request.form:\r\n image_url = request.form['image_url']\r\n response = requests.get(\r\n '%s?image_url=%s' % (ENDPOINT, image_url),\r\n auth=auth)\r\n\r\n try:\r\n res = response.json()\r\n except Exception as e:\r\n print('Exception in JSON decode:')\r\n print(e)\r\n print(response.content, response.status_code)\r\n\r\n return render_template('index.html', image_url=image_url, res=res)\r\n\r\n\r\nif __name__ == '__main__':\r\n app.run(debug=True)\r\n<\/pre>\nEvery call to the Imagga API must be authenticated. Currently, the only supported method for authentication is\u00a0Basic<\/a>. With Basic Auth, credentials are transmitted as user ID\/password pairs, encoded using base64. In the above code, we achieve this with a call to\u00a0
HTTPBasicAuth()<\/code>.<\/p>\nWe then create a function that will be triggered by GET and POST requests to the\u00a0
\/<\/code>\u00a0route. If the request is a POST, we get the data submitted by form and send it to the Imagga API for classification.<\/p>\nThe NSFW Categorizer is one of a few\u00a0categorizers<\/a>\u00a0made available by the Imagga API. A Categorizer is used to recognize various objects and concepts. There are a couple predefined ones available (Personal Photos<\/a>\u00a0and\u00a0NSFW Beta<\/a>) but if none of them fit your needs we can\u00a0build a custom one for you<\/a>.<\/p>\n
As mentioned previously, to send an image for classification, you send a GET request to the
\/categories\/<categorizer_id><\/code> endpoint. Thecategorizer_id<\/code> for the NSFW API isnsfw_beta<\/code>. You can send the following parameters with the request:<\/p>\n- \n
- image_url<\/strong>: URL of an image to submit for categorization.<\/li>\n
- image_upload_id<\/strong>: You can also directly send image files for categorization by uploading the images to our
\/uploads<\/code> endpoint and then provide the received content identifiers via this parameter.<\/li>\n- language<\/strong>: If you\u2019d like to get a translation of the tags in other languages, you should use the language parameter. Its value should be the code of the language you\u2019d like to receive tags in. You can apply this parameter multiple times to request tags translated in several languages. See all available languages\u00a0here<\/a>.<\/li>\n<\/ul>\n
After processing the request, the API sends back a JSON object holding the image\u2019s categorization data in case of a successful processing, and an error message incase there was a problem processing the image.<\/p>\n
Below you can see the response of a successful categorization:<\/p>\n
{\r\n \"result\": {\r\n \"categories\": [\r\n {\r\n \"confidence\": 99.1496658325195,\r\n \"name\": {\r\n \"en\": \"safe\"\r\n }\r\n }\r\n ]\r\n },\r\n \"status\": {\r\n \"text\": \"\",\r\n \"type\": \"success\"\r\n }\r\n}<\/pre>\n<\/p>\n
Note that you might not always get JSON with the three categories displayed. If the confidence of a category is\u00a0
0<\/code>, this category will not be included in the JSON object.<\/p>\nBelow you can see the response of a failed categorization.<\/p>\n
{\r\n \"status\": {\r\n \"text\": \"Unexpected error while running the classification job.\",\r\n \"type\": \"error\"\r\n }\r\n}<\/pre>\nBack to our app, you can save your code and run it with:<\/p>\n
\r\n$ python app.py\r\n<\/pre><\/p>\n
- image_upload_id<\/strong>: You can also directly send image files for categorization by uploading the images to our
- underwear<\/strong>\u00a0– this categorizes medium safe images. These might be images displaying lingerie, underwear, swimwear, e.t.c.<\/li>\n
- Community driven<\/strong>\u00a0– with this type of moderation, the owner of the application puts in place features that enable the app\u2019s users to report any inappropriate content e.g. flagging the content. After a user flags a post, an admin will then be notified. This also suffers from a delay in identifying inappropriate content from both the community (who might not act immediately the content is posted) and the administrators (who might be slow to respond to flagged content). Leaving moderation up to the community might also result in reported false positives as content that is safe is seen by some users as inappropriate. With a large community, you will always have differing opinions, and because many people will probably not have read the Terms and Conditions of the medium, they will not have clear-cut rules of what is and isn\u2019t okay.<\/li>\n
![\"\"](\"https:\/\/imagga.com\/blog\/wp-content\/uploads\/2018\/01\/CAR-800x450.jpg\")
<\/p>\n![\"\"](\"https:\/\/imagga.com\/blog\/wp-content\/uploads\/2018\/01\/PEOPLE-800x450.jpg\")
<\/p>\n![\"\"](\"https:\/\/imagga.com\/blog\/wp-content\/uploads\/2018\/01\/Underwear-800x450.jpg\")
<\/p>\n