OICO Research and Development projects
1- Digital Eye Test Chart.This was the first product we developed four years ago and was a simple "taster" project to begin applying our skills to the medical device area. It was fundamentally a simple project and we designed it from first principles for simplicity, ease of use and low cost. The final concept became a small ARM embedded computer as the main device, connected by its own wifi acting in Access Point mode to a low cost android tablet as the controller for the medical professional. Our device brought several unique aspects for this product category including the first user interface that took full advantage of modern capacitive touch screens, eliminated sub menus from the user interface, thus enabling the user to get up and running with the system extremely quickly. By using a small ARM microcomputer for the main test chart we also eliminated the need for a desktop PC controller. Overall we achieved a cost reduction of 75% compared to incumbent digital test charts in the market as well as improved usability and reliability through the use of low-cost solid state components. We developed the entire software (both the ARM and the Android) using rapid development tools.
Current status: In production for over 3 years, over 2 million patients in the UK, Colombia, Brazil, Ireland, Nigeria and other countries have been tested with this device. Failure rate >2%, return rate >1%
Further information can be found here
2- Low Vision Book Reader.This was the second project we undertook at OICO. The idea was to create a low cost device that enabled people with low vision to access books and was developed at the request of customers of our first device. Going back to first principles, we took, for the first time, an agile approach to this product development. With the product owner and my small team we defined the initial specifications of what the end user needed and settled on a simple initial design which incorporated an ARM micro computer with several sensors (camera for producing enlarged text on a screen, microphone for voice commands and speakers for reading out books for the blind), we sat down with low vision patients and evaluated their abilities to interact with technology. On the basis of this evaluation we added a joypad controller to the system which is wired (to avoid misplacement of joypad by the end user as well as the need to recharge). At every point we considered the factors that affect the usability of the device for the end user. We found that the available offline voice recognition systems were not accurate enough for frustration free use which is why we ended up designing a system that uses either a physical joypad controller or is connected to web service based voice recognition systems. Software was developed with python and for the enclosure we investigated different potential case shapes, materials and colours to make it both easy to find for low vision users as well as providing the right acoustic qualities for clear text to speech reproduction. The joypad user interface, text to speech book reader and screen enlarger were fully developed, however due to the then imminent release of the Amazon Echo it was decided that our new product was not competitive and was discontinued before production commenced. A single example was assembled, with a wooden case for better acoustic production and a Braille enhanced joypad for control.
Current Status: 1 prototype produced. Release of Amazon Echo made product obsolete before reaching market.
3- Digital Fundus Camera.With desktop fundus cameras costing from £8000-£15,000 per unit, their costs were prohibitive for many developing world users who did not have access to the easy low rate finance that UK and European based medical centres and opticians had access to. Thus we developed a new digital fundus camera from first principles taking just the expired original patents for a general cursory working principle, but beyond that designed the unit from scratch to massively cut cost and development time as well as improve usability and reliability. The optics, mechanical components, microcontroller for optics / motors / photonics as well as the user interface software (python / DICOM as well as some small C based vision programs) were all created in house. The device reduced a fundus camera from a part count of 700+ (we bought a competing camera unit and took it apart to see this) to 40 parts. The entire photonics system was reduced to a single 3D printed part into which photonics and filters and LEDs slotted in. Optical redesign reduced optics paths and the use of simple Python programming enabled us to roll out units within a few months of commencement.
Current status: The product proved the design principles needed to make an effective non-mydriatic fundus camera using a brand new additive manufactured design that reduced part counts by 90% and overall bill of material price by nearly 70% compared to the competition. However customers in the developed world did not want a desktop unit as they mostly had one already, and in the developing world they found the cost to still be too high and the design did not meet their needs in operating in very primitive conditions and in mobile clinics.
4- Portable Fundus Camera.Capitalising on the optics and basic design of our desktop fundus camera, we took on board the concerns of potential clients and redeveloped the unit to be lower cost and hand portable. This wasn't a massive task as we quickly settled on the use of a Smartphone which already has the camera sensor, battery, screen and processor and incorporated it into a slightly redesigned system from the desktop mounted camera. The design, naturally made modifications to the camera to expand its spectral sensitivity as well as change the focusing mechanism and focal range, but beyond that the optics remained the same. In terms of software, the desktop oriented Python+C software was rewritten for Android and the user interface changed drastically from a mouse and keyboard setup to a design optimised for capacitive touch screens. The unit was of course 3D printed with part count kept to 40 units.
Current Status: In production for clients in the UK and abroad. Some clients requested that the unit should become more stable to take better pictures as the pupil diameter can be as small as 2-3mm diameter with both the patients head as well as the medics' hand moving slightly. This brought us onto the product 5.
Further information can be found here.
5- Portable desk top mount.Developed for the portable fundus camera At the behest of customers and keeping in line with our standard practice of agile development we looked at our and our competitors portable fundus camera offerings. Seeing that hand movement of the operator and head movement of the patient made a major degradation in the image quality obtainable we looked at ways of overcoming that issue. After deliberating over the use of software and electronic stabilisation of the sensor it was deemed insufficient to improve quality and steady head and device were paramount. So we designed a portable mount for both the patients head as well as a mount for the portable camera to be mounted directly on the carry case which holds the camera and thus not increasing the package size for portability and increasing the overall weight by less than 500 grammes. The new mount was designed to offer the same flexibility by giving three degrees of freedom in adjustment to obtain stable line of sight between patient pupil and the camera sensor and achieving in a low cost compact package an image quality that is similar to 4-5 year old desktop mount cameras (sensor limitations mean that it cannot match the latest high Quantum Efficiency desktop cameras).
Current Status: In exclusive negotiations with a large domiciliary Optical company in the UK for use with OICO portable fundus cameras
6- Portable AGE Test.Portable Advanced Glycation End Product device for early symptoms of possible onset of Diabetes. Working from the principles outlined in academic works on the subject of AGEs I led a team of two people (myself in charge of software, electronics, enclosure and my colleague in charge of Optics, Photonics and mathematical models) to develop an AGE reader and associated software. The AGE reader was developed in an Agile development process and fully utilised additive manufacturing to massively reduce part count and complexity of assembly. We developed several new methods for cost reduction including the elimination of the use of expensive Single Photon Counters (APD and Silicon Photo Multiplier Tubes) and instead made use of low cost photodiodes with associated bandpass filters for specific wavelengths to achieve acceptable performance for this limited role (through the utilisation of a 90%+ QE photodiode for that wavelength as well as high efficiency filters. This also enabled a massive reduction in the design complexity, power usage dramatically reducing the overall complexity of the design. The bill of material cost is 10% of an equivalent APD. Current Status: In Development. Software ASSAY database, calibration and QA for diagnosis is being developed
7- Multi Spectral TomographyMulti Spectral Tomography System as a low cost and portable alternative to Optical Coherence Tomography. At OICO we had a number of users of our Portable Fundus Camera ask us if we had developed an OCT (Optical Coherence Tomography). We investigated the idea and went back to first principles of what was sought by the end user. An ability to image the tissue under the surface of the retina. After much deliberation and prototyping we come to the conclusion that we could use images taken at different frequencies and construct a depth image far more cost effectively and simply than the use of a single tomography unit. We designed and built a prototyped around a USB3 vision imaging sensor and a broad range of LEDs for illuminating the retina at different frequencies. We designed the 3D printed enclosure for prototyping and built the embedded ( C ) as well as front end (C#/WPF and ImageJ) software to enable testing and optimisation to begin. Right now image and section data optimisation is being undertaken by experts who benchmark against OCT units.
Current Status: Prototype in development