SP's: penFace, the open source face recognition technology developed by the researchers at the Carnegie Mellon University, has been rolled out. The OpenFace holds ground against the much-advanced proprietary face-recognition technologies by private endeavours.
Open-source technology is usually at par with the proprietary technology but in certain cases such as the face-recognition, the open-source lags far behind.
The intelligentsia of the open-source community has now stepped up and created an open source face recognition system. The researchers at Carnegie Mellon University feel that the accuracies of the open-source face recognition system lag behind the advancements garnered by the tech giants with their exclusive state-of-the-art.
Open source, as the name gives away, is the technology that derives its ideas from the 1970’s hacker culture defining the craze for programming and its subsequent availability to all. This philosophy was later subsumed by FOSS (Free and Open Source). The lack of resources for the projects that include hardware technology sometimes hinder the excellence in the open source community.
OpenFace is based on Google’s FaceNet face recognition system and can recognise the faces within few seconds with 10 reference photos. The datasets that the developers have used to train this program to recognize faces falls to a fraction as compared to the huge datasets of private endeavours.
However, the developers are quite happy with the results as the accuracy they achieved is remarkably high and exceeds in performance in comparison to the previous open source face recognition technology attempts on the standard LFW benchmark.
Watch OpenFace in action, recognizing the faces in real-time:
If you belong to an open-source community, perhaps you should take a look at the OpenFace project.
For more information on the OpenFace project, visit GitHub.com
Post a Comment