{"id":22028,"date":"2020-03-24T17:44:40","date_gmt":"2020-03-24T17:44:40","guid":{"rendered":"http:\/\/dam_for_developers_ai_based_content_analysis"},"modified":"2024-06-05T16:39:05","modified_gmt":"2024-06-05T23:39:05","slug":"dam_for_developers_ai_based_content_analysis","status":"publish","type":"post","link":"https:\/\/cloudinary.com\/blog\/dam_for_developers_ai_based_content_analysis","title":{"rendered":"DAM for Developers: AI-Based Content Analysis"},"content":{"rendered":"

Content analysis on images based on artificial intelligence (AI) is yet another powerful feature of Cloudinary\u2019s digital asset management (DAM) solution<\/a>. This article explains the types of analysis Cloudinary DAM can perform for you.<\/p>\n

To run an analysis on an image in your Media Library, click the image from the dashboard and then the Manage<\/strong> button near the top right corner. A pane then opens on the left with three tabs at the top: Summary<\/strong>, Metadata<\/strong>, and Analysis<\/strong>, as shown below:<\/p>\n

\"Analysis\nAnalysis Tab for an Image<\/em><\/p>\n

Click the Analysis<\/strong> tab for the related display:\n\"ai<\/p>\n

Cloudinary already runs some basic, AI-based analysis on the images you\u2019ve uploaded to your Cloudinary account\u2019s Media Library. As the above screenshot shows, you must register for add-ons before certain types of analysis can occur.<\/p>\n

Following are the analyses you can run for your images with Cloudinary DAM.<\/p>\n

Auto-Tagging<\/h2>\n

Thanks to deep-learning AI models, which categorize images\u2019 object types, you can automate the process of marking each of those categories as tags for your images.<\/p>\n

First, click Register<\/strong> under Free Registration for Amazon Rekognition<\/strong>, as shown here:\n\"Amazon<\/p>\n

Now run the analysis to show the objects in the image. See this example:\n\"Amazon<\/p>\n

You can then assign all those categories as tags to your images. What a time-saver!<\/p>\n

Detection of Faces and Celebrities<\/h2>\n

\"Face<\/p>\n

By leveraging AI\u2019s face-detection algorithms, Cloudinary can name the celebrities in an uploaded image. Each face- or celebrity-detection option displays in the preview pane on the right\u2014in color-coded frames\u2014the coordinates of the faces it finds. To show or hide those frames, click the Show<\/strong> or Hide<\/strong> button.<\/p>\n

Also displayed is the accuracy and confidence level of the celebrities in the image.<\/p>\n

To activate this feature, click Register<\/strong> under Free Registration for Amazon Rekognition: Celebrity Detection<\/strong>, like this:<\/p>\n

\"Celebrity-Detection\"\nActivation of Celebrity-Detection Add-On<\/em><\/p>\n

Now run the analysis to find out who the celebrity in the image is.<\/p>\n

\"President\nDetection of Celebrity<\/em><\/p>\n

Sure enough, Cloudinary detected the celebrity as Barack Obama with 100-percent accuracy. Bravo!<\/p>\n

Detection of Optical Characters<\/h2>\n

Cloudinary DAM can read the text in uploaded images by running an analysis of optical-character recognition (OCR) on them.<\/p>\n

\"OCR\"\nImage With Text Before Cloudinary DAM\u2019s OCR Analysis<\/em><\/p>\n

To activate that feature, click Register<\/strong> under Free Registration for OCR Optical Character Recognition<\/strong>. Afterwards, run the OCR analysis. Here\u2019s an example of the result:<\/p>\n

\"text<\/p>\n

Color Analysis<\/h2>\n

Subsequent to a color analysis that yields information about the main colors in an image, Cloudinary DAM can automatically add background padding to it in an attractive color.<\/p>\n

Additionally, with the premium feature Image Analysis<\/a>, an advanced Media Library search capability, you can search for images based on their predominant colors.<\/p>\n

\"Color\nColor Analysis of an Image on Cloudinary DAM<\/em><\/p>\n

Conclusion<\/h2>\n

Cloudinary DAM\u2019s AI-based content analysis capability serves innumerable use cases for developers. Do share the creative ways you\u2019re applying it to your projects. We can all learn from one another\u2019s ideas. Check out my next post in the DAM for developers series about access control<\/a>.<\/p>\n

\n

DAM for Developers: Series<\/h3>\n