Voxeleron LLC was co-founded in 2010 by Daniel Russakoff and Jonathan Oakley to develop state-of-the-art software in the field of computer vision and machine learning. Our specialty is in the development of ophthalmic image analysis software for optical coherence tomography (OCT) imagery. In this area, we have unrivaled experience in developing and releasing analysis algorithms. We have profound expertise in computer vision, machine (and deep) learning and ophthalmic imaging. Most organizations struggle to find expertise in any two of these areas. This has resulted in the release of Orion, the first commercially released, device-independent OCT analysis software, and also InSight, our animal OCT application.

Consultancy services specializing in medical image processing.

Research, development and implementation of algorithms for Optical Coherence Tomography (OCT) to support disease management and diagnosis. This includes methods for volumetric segmentation, image registration and motion correction. Responsible for the initiation of new research projects as well as supporting improvements to existing products. Support for the development of new clinical applications, the meeting of market requirements, the generation of scientific publications and the securing of intellectual property rights. At CZMI my algorithms that I put into product include: Cirrus motion correction, all Cirrus registration algorithms, retinal layer segmentation, optic disc centering (AutoCenterTM), corneal segmentation, Selective Pixel ProfilingTM, and the optic nerve head analysis algorithms. The development of inner retina layer segmentation algorithms at an accuracy unseen in existing software spawned numerous clinical findings that resulted in seminal journal publications in Brain, Neurology, Ophthalmology etc. It put new technologies in the hands of clinicians, who underscored the importance of such inner retinal layers as they relate to neurological diseases.

Responsible for the development and deployment of the machine learning algorithms used for fingerprint "liveness" recognition. Given the excellent real-world performance of these so-called spoof detection algorithms, they became a key technology in the Lumidigm portfolio that was eventually acquired by HID Global. Additional projects included the keypoint detection and matching, but my focus was on feature detection and classification algorithms used for machine learning and pattern recognition.

See other description for work at CZMI.

Development of computer-aided detection and classification software for medical imaging applications. Research and development of new algorithms for image pre-processing, object recognition, feature extraction, shape description and classification. Responsible for the release management of the mammography CAD code, and initiated strict quality control procedures to improve the coding standards.

Responsible for the product development and characterization of semiconductor inspection tools and their pattern recognition algorithms. Duties ranged from the scoping of the product’s technical feasibility, through to the product’s Beta implementation and acceptance at the customer’s site. Membership of both the algorithms and the ease of use task forces responsible for defining and prioritizing future R&D initiatives. Additionally, I received the General Manager award for outstanding contributions to the division in fiscal year 2003.

I was a research scientist at the PSI during my PhD studies at ETHZ.

Relational database programmer. This was an internship - or sandwich year - during my studies as a Computer Scientist at the University of York.