Watermark detection

A ML model to identify the presence of visual watermarks in input images

Overview

Protection of intellectual property and brand integrity is of great importance to content creators, marketers and organizations. Watermarks, which are often applied to images to prevent unauthorized use, play a significant role in safeguarding digital assets.

The Watermark detection functionality has been developed as a crucial component of our service. This machine learning model is designed to accurately determine whether an input image contains a visual watermark, thus providing customers with the ability to identify and work with their digital assets more effectively.

The employed deep learning algorithms analyze the input image data, extracting unique features and patterns that are indicative of visible watermarks, with the end goal being distinguishing between watermarked and non-watermarked images with high accuracy.

Typical use cases

Watermark detection can be employed in a variety of use cases, including:

• Licensing compliance - Media agencies can use the feature to ensure that content consumers adhere to licensing agreements, thus preventing unauthorized use of their assets.

• Asset management - The model can automatically categorize and tag images, allowing users to quickly locate and retrieve non-watermarked versions of their assets.

• Quality control - Design and marketing teams can easily verify that images in promotional materials are free from unintended watermarks, ensuring consistent branding.

API endpoints

Information about the specific API endpoints is available in an always up-to-date documentation, that can be accessed via the following link:

Scaleflex API for Digital Asset Management (DAM), Visual AI and Media OptimizationScaleflex API

There, you can find detailed information about the API endpoints, together with all required request parameters, so you know how to interact with them.

Example API responses

Input image	Input image

API response	API response
{ "status": "success", "version": "3.0.2", "predictions": [ { "no_watermark": 0.95, "watermark": 0.05 } ], "request_uuid": "6d74d4b0-c761-4083-8e6d-db03de49065e", "sha1": "3ac8ffca25a63144ca14cd493449a69d2442fb30" }	{ "status": "success", "version": "3.0.2", "predictions": [ { "no_watermark": 0.1, "watermark": 0.9 } ], "request_uuid": "d9fab3aa-e8be-4d6b-a225-4d530b75416f", "sha1": "6db8da981658f926c8cc435fd378cc6d307463d3" }

Input image	Input image

API response	API response
{ "status": "success", "version": "3.0.2", "predictions": [ { "no_watermark": 0.949, "watermark": 0.051 } ], "request_uuid": "eb83ecf4-16e2-4f02-bfd6-b5059d760675", "sha1": "d15bb19c4b0081a564c298ad66b583b2f0f405fa" }	{ "status": "success", "version": "3.0.2", "predictions": [ { "no_watermark": 0.189, "watermark": 0.811 } ], "request_uuid": "608f91e3-2e24-430e-963d-5ae6ee211084", "sha1": "e0af32948b6469a8bef58ece415b38ed392595a5" }

Input image	Input image

API response	API response
"status": "success", "version": "3.0.2", "predictions": [ { "no_watermark": 0.973, "watermark": 0.027 } ], "request_uuid": "ae15f794-e98c-4f34-a75b-85de56d82e5a", "sha1": "6a5614e25e08663ebb1418c6d3090476b5bc8941" }	{ "status": "success", "version": "3.0.2", "predictions": [ { "no_watermark": 0.083, "watermark": 0.917 } ], "request_uuid": "5d7a8ce6-a3e0-47cc-b587-496465b5256e", "sha1": "c1c3cf2effd2d822691858c32335b72c7ff9e8a7" }