Photo Authenticity
Streamline Your Processes
Working with photos from third-party sources can be difficult when the legitimacy of pictures is in question. We have developed several tools to verify the validity of third-party images so you can securely operate your business.
Photocert is continuously researching and developing tools to detect and localise fraud within JPEG and PNG images. The current tools for image verification include:
General Analysis
Different physical characteristics of the media file are analysed and compared to market standards in order to assess the likelihood of the photo being tampered with. This initial analysis also helps in understanding the effectiveness of other tests.
Image Trace Detection
If a picture has been modified using photo editing software, our computer vision and machine learning algorithm can detect traces left on the photo and determine which part of the photo has been edited.
Web Search Analysis
Photo web search is a great tool to find if a picture (whole or partially) can be found on the web. The file is compared using an accurate reverse photo search that can also discover the image in different sizes and crops of the original image.
Meta Data Check
Metadata is an essential part of a file, embedding detailed information about the creation and modification of a file. The analysis of the metadata contained in a picture can provide several insights.
Similarity Search
A search within an internal client database avoids the risk of receiving photos that have already been used. Newly uploaded pictures are processed and compared with existing ones to establish whether there are any similarities.
Picture-of-a-Picture Analysis
Image taken from a computer screen.
Multiple analyses are employed to assess the authenticity of photographs. However, a distinct challenge arises when the image itself is genuinely authentic, yet the information it conveys is deceptive. Such a scenario is possible in the context of a “picture-of-a-picture,” wherein an authentic photo is captured from a printed document or a screen, enabling the user to transit false information.
Photocert has developed AI analysis capable of detecting photos that have originated from printed documents or screens, thereby preventing the possibility of bypassing other authenticity analyses.
Generative AI Identification
Image fully generated by an AI.
In today’s rapidly evolving technological landscape, the rise of generative AI has introduced a new set of challenges. As AI-driven content generation becomes increasingly sophisticated, the potential for misuse and misinformation grows. It has become essential to address the emerging problem of identifying AI-generated content to maintain trust, security, and data integrity.
Utilizing proprietary AI-powered analysis, Photocert is capable of identifying fully AI generated photos from the most widely used platforms, while also detecting photos that have been partially generated using AI.