Featuring comprehensive full and segmental image search capabilities, using your own cloud domain library, comparative 1:1 COMP-V1, along with VidG5 for videos.
Release : November 4, 2026
We are considering the scalability of this product. If you are an investor interested in allocating resources towards valuable development initiatives, we invite you to connect with us at contact.general@keras360.io, or may contact our founder, the primary architect behind the protocols, at nylin@naorococo.net.
We are pleased that you are following our updates on Art Forensics. In an era characterized by advancements in technology, where digital art is transforming from traditional film negatives to dynamic digital imagery, Keras360 is at the forefront of reimagining the potential within data science. You have observed the progress of our beta experiments that focus on the recognition of portraits through comprehensive segmentation techniques applied to our image library (ART-Z : AF library), which has laid the groundwork for the development of our AF Basic user interface (UI).
Our Mid-Range AF UI now facilitates the launch of an image editor, enabling the editing of uploaded images prior to the authentication process. Keras360, along with our analysts, has thoroughly considered the importance of protecting individuals’ privacy in the realm of image analysis. Consequently, in addition to our existing image collection such as test libraries: ART-Z for portraits and ART-MOV for image frames from videos, users are empowered to upload their image of interest and conduct an automated analysis against their library stored in their Cloud-Drive.
Aside from personal cloud domain library search, this prototype release marks the inaugural presentation of our protocol Comparative-V1 aka. COMP-V1 (2024/2025, stable); a comparative protocol designed for the upload of two images: one test image and one reference image to be submitted by the end-user; refer to Video 2 below.
Image analysis protocols in Mid-Range UI
Requires subscription and login credentials
Full access
In the Keras360 demonstrations, we avoid showing the navigation through our folders to find images for upload. The video features segments edited to exclude the image search. Please concentrate on how the analysis progresses to achieve a conclusive outcome.
Video 1 : Art Portraits
♠ Find-Match (Version 2): The reference image is contingent upon the cloud-library collection while the user uploads a test image. The agent application endeavors to identify a match within the library. Should a match be located and the test image exhibit authentic characteristics, AF will return the match profile.
Find-Match : Cloud Dependent
It is important to note that the preference for an increase in sensitivity detection will prompt a validation process during the Art Portraits theme search.
🎵 Music by : Ryo Yoshimata, Sound of Ocean
Video 2 : COMP-V1 vs Cloud-Dependent Search
♠ COMP-V1 (2024/2025, stable): Test and reference images are uploaded by the user.
Video below shows two protocols: COMP-V1 and Ceramics Collection cloud-dependent search
– Applicable for ceramics and mint specimens.
– COMP-V1 (0:00)
– Ceramics test-cloud data library (0:12)
Some may contend that the aforementioned search does not adequately substantiate the utility of the COMP-V1 protocol. It may be prudent to review the part 2 video: https://keras360.io/art-forensics-ui/mid-range-af/
Images origin : Christie’s website
Video 3 : Be@rBrick + VidG5 video splicing and image cloud-dependent search
♠ Video-Splicing (July 2026): The initial five seconds of the video upload are extracted, and up to five selected image frames are processed concurrently during the authentication procedure.
Video-Splicing Feature: The uploaded video has a duration of two seconds. ART-MOV serves as a test cloud data library for moving images and contains split images of the bearbrick. View the actual video footage located at end page : https://keras360.io/art-forensics-ui/mid-range-af/.
Upon the completion of the image frame extraction, the user has the opportunity to select the preferred search domain; utilizing the art-portraits theme, a personalized cloud data library, or any of the protocols within the mint or ceramics & porcelains categories.
📹 👀 Jump to section in Video 3
– Video to Image frame extraction (0:10)
– Orientate images (0:17 )
– Select images (0:26)
– Analyze image frames using default protocol (0:31)
– Use extracted image as a portrait search (0:40)
– Personalised data library (0:56)
Available themes and Cloud collections
♠ Art Libraries
(ART-Z; aka. AF library, etc. )
♠ Mint Collections (Banknotes, Coins)
♠ Ceramics and Porcelains
♠ User’s Cloud Image Library
Learn about the UI features :
Earlier protocol developments :
