CheckDIY Images

Pui Ling Hon

In-app purchases

About this app

This app equips you with both manual analytical techniques and machine learning models to examine images and determine whether they have been altered.

The app incorporates multiple forensic techniques, including Error Level Analysis (ELA), Noise Analysis, Local Binary Pattern (LBP), Copy-Move Detection, and JPEG Ghost Detection. Each technique focuses on different aspects of an image, helping to uncover potential tampering. In addition, the classification machine learning model identifies whether an image has been altered, while the region-based model highlights areas where modifications may have occurred.

It is important to note that this app is designed to detect MANUAL IMAGE MANIPULATION, rather than AI-generated alterations. It is most effective for detecting edits made with traditional tools like Photoshop, where image layers, content-aware fill, or cloning techniques have been used. However, it may NOT accurately detect images that have been GENERATED OR MODIFIED BY AI MODELS (such as deepfakes or AI upscaling).

Since no single technique is foolproof, it is strongly recommended to apply multiple forensic analysis methods and machine learning models to get the most reliable results. Relying on just one method may not provide an accurate conclusion. This app is a tool to assist with image analysis, but it should be used alongside other forensic methods for a more complete assessment. It is not a substitute for professional forensic examination.

What’s new

This release update the app to build in the recent target api, 35.