- We began rolling out an improvement to our deep learning algorithms—this effects both our cloud and on-premises versions of the Kairos API. This upgrade significantly improves face-detection and matching accuracy for all faces. It's an automatic update, so there's nothing for you to do.
- We’ve been working hard on an upgrade to make Kairos’ API faster and more stable for you —dramatically improving overall performance. This upgrade fixes issues with slow response times and increased error rates when making API calls.
- We shipped a huge update to Kairos’ algorithms; we’re now 100% Deep Learning, improving accuracy and performance for all customers of the API.
- We launched an On-Premises version of our popular Face Recognition Cloud API; customers can now self-host a version of Kairos in their own datacenter.
- We have a big enhancement being released today. We have rewritten our entire facial recognition backend in Go!
- What does this mean for you? Not much other than expect to see much more consistent results, and vastly improved response times across all API calls!
- We also added two new parameters, "show_uploaded_image" which when set to "true" will include a 1 hour temporary URL where you can retrieve the image that was just sent in for the /enroll call.
- The other parameter that was added is called "show_candidate_images" which when set to "true" will send you a temporary URL for each candidate that appears in a /recognize call. This was a hotly requested feature that we are glad to be able to deliver for you.
- Note: We did correct an issue with the /enroll call where the "face_id" parameter was returning an integer and not the correct face_id corresponding to what appears in the /recognize call. This has been corrected and a string is now returned.
- Based on your feedback, we have added the uploaded_image_url to the /verify, /enroll, and /recognize methods!
- Note that this URL is only good for one hour after it has been created.
- Look for more new features in this area coming very soon!
- We have added the eyeDistance metric to the /verify, /enroll, and /recognize methods!
- New major enhancement to the accuracy of the facial recognition engine
- You should notice a big speed increase for repeated /recognize requests to the same gallery
- You asked for it, we listened. You can now remove a single face from a subject_id without removing the entire subject!
- Check out the new View subject method to see the face IDs enrolled for a particular subject
- You can now pass in a face_id to the Remove subject method to remove just one face from a subject
- The face_id is now passed back to you when you issues an Enroll request
- We have added a "maleConfidence" and "femaleConfidence" values to the /detect method
- Added a detector for lips (Together, or Apart) to our face recognition API
- Added a detector for glasses (Sun, Eye or None) to our face recognition API
- More improvements to speed when performing recognize operations on very large galleries.
- Proud to announce some big upgrades today to our face recognition API.
- Ethnicity detection added to /enroll and /detect methods.
- Improved performance in both accuracy and speed of /recognize and /verify operations.
- Added a new face "quality" metric to let you know how well your image will perform for later recognition or verification.
- We have added a new "form upload" capability to the face recogition API in addition to passing in images as URLs and Base64 encoded.
- We have made some performance enhancements to our face recognition API. You should see dramatic improvements in speed when searching very large galleries.
- Added a completely revamped and improved high performance face recognition engine available to all users.
- Verify method now available for all face recognition users.
- Age and Gender detection now available for all face recognition users.
- Multi-face analysis added to our Emotion API.
- Released version 2 of our Emotion API now with age, gender, 6 emotions and more.
- API examples updated with improved visualizations for our Emotion API.
- One to one face recognition verification is now available with our /verify method for our Enterprise Plan customers. See the docs here.
- Check out our new API examples for our Emotion API. Watch this space for more to come soon!
- We have opened up the source for our Face recognition Android client library . Pull requests are welcome!
- We have opened up the source for our Face recognition iOS client library . Pull requests are welcome!
- Emotion API client libraries are now available
- We have released our new Emotion API. You are able to receive detailed facial analysis on videos sent into our new /media endpoints. See our new Emotion API documentation site.
- Added new gallery/remove method to the API and SDKs which will enable you to remove all the subjects in a gallery with one call and remove the gallery itself
- Android SDK now available
- Standardized error codes
- New API documentation site released
- Gender detection now available in all face methods
- PHP SDK now available
- CORS (Cross-Origin Resource Sharing) support added to the API
- We now automatically rotate incoming photos
- Added a parameter for minHeadScale to enable detection of images where the heads take up a small portion of the overall image
- Added ability to process multiple faces during enroll
- Added ability to accept and auto-detect a URL or Base64 image using the image parameter
- url parameter deprecated
- Added selector parameter to enroll and recognize methods
- Added symmetricFill parameter for all face methods
- SSL support now available at https://api.kairos.com
- iOS SDK available
- Second Ruby gem wrapper for Kairos ID available
- New lower pricing and increased API call limits
- New workflow for enroll and recognize, no longer need to call detect prior
- New gallery methods added for managing your galleries
- Base64 encoded images can now be submitted to the API
- Kairos ID generally available
Detect, enroll and recognize methods for 2D to 3D facial recognition