Image-Based OSINT Investigations: Tips & Techniques

In the world of Open Source Intelligence (OSINT), images can be a goldmine of valuable information. They can provide visual evidence of a subject’s appearance, their whereabouts, and even the vehicles they use. By leveraging search engines and free tools, investigators can extract valuable intelligence from images, such as identifying the location and time the photo was taken, the devices used, and even connecting social media accounts to a subject. In this article, we will explore the techniques and tools for conducting image-based OSINT investigations, including reverse image searching, facial recognition, deepfakes, and metadata analysis.

Reverse Image Searching

Reverse image searching is a powerful technique that allows investigators to discover visually similar photos from around the web. By uploading an image or inputting its URL, you can ask a search engine to locate and display related images used on other websites. This technology, known as content-based image retrieval (CBIR), enables investigators to identify related images based on various factors, such as similar visual content, objects, or people.

Reverse image searching can be particularly useful in investigations involving statues, buildings, places, people, and logos. By recognising a statue or building in the background of an image, search engines can help identify the location where the photo was taken. Additionally, search engines can locate other images of your subject or logos on websites that provide relevant information.

There are several search engines and dedicated reverse image search sites that you can utilise for this purpose:

1. Bing Visual Search: Great for flipped and altered images, and facial recognition.
2. Yandex Visual Search: Ideal for analysing faces, buildings, and locations.
3. Google Images Search by Picture: Useful for identifying buildings, locations, and logos.
4. Tineye: Effective for identifying logos and alternate versions of the same image.
5. PimEyes: Excellent for facial recognition purposes.

Facial Recognition

Facial recognition systems utilise AI algorithms to detect, identify, and analyse faces in images and videos. By analysing facial features, such as eyes, nose, mouth, and face shape, these systems can create a mathematical formula that represents a person’s face. This formula can then be compared with other images to determine if there is a facial match.

Two notable facial recognition tools are worth mentioning:

1. Microsoft Azure’s Face Verification: This tool enables users to determine the likelihood that two uploaded images containing faces belong to the same person, expressed through a confidence score.
2. Amazon Rekognition: Amazon Rekognition can detect faces in images and videos, providing information about where faces are detected, facial landmarks, and detected emotions. It also allows for face comparison between images.

Facial analysis can be performed on uploaded photographs to extract additional details about the individuals in the image. Rekognition’s face comparison feature provides a confidence score indicating whether the persons in two photos are the same.

Deepfakes

Deepfakes are realistic simulations of audio and video media created using deep learning algorithms. These techniques can be used to make media that depicts fake events or manipulates the appearance and actions of individuals. Deepfakes have the potential to deceive people by showing politicians or celebrities saying or doing things they never actually did. They can also be used for criminal purposes, such as replicating voices to access bank accounts or convincing individuals to transfer money.

In the realm of OSINT, it is essential to be able to identify, verify, and debunk deepfakes. Fortunately, tools like the Deepware Scanner can help in this regard. The Deepware Scanner is specifically designed to analyse suspicious videos and identify if they have been synthetically manipulated.

Metadata Analysis

Metadata is data that provides information about other data, such as the time and location an image was taken or the camera used. While many social media sites remove metadata from uploaded images, original digital photos often retain this information. Analysing metadata can provide valuable insights into the origin and context of an image.

One tool that lets you view metadata is Jeffrey’s Image Metadata Viewer. This browser-based tool allows you to upload a photo and view its EXIF data, which includes details like the time and date the image was captured, the camera model, and even the location if enabled on the camera.

For a comprehensive analysis of metadata, including camera details, shutter speed, and embedded coordinates, Jeffrey’s Image Metadata Viewer is a valuable resource.

Conclusion

Image-based OSINT investigations offer valuable insights into subjects, locations, and events. By utilising reverse image searching, facial recognition, deepfake detection, and metadata analysis, investigators can enhance their intelligence gathering capabilities. These techniques provide a comprehensive solution for conducting effective image-based investigations. Embrace the power of images and unlock a wealth of information in your OSINT endeavours.