Mobile Search for the Visually Impaired

DESCRIPTION (provided by applicant): The technology developed as part of this NIH SBIR project will transform the cell phone camera of visually impaired individuals into a powerful tool capable of identifying the objects they encounter, track the itemsthey own, or navigate complex new environments. Broad access to low-cost visual intelligence technologies developed in this project will improve the independence and capabilities of the visually impaired. There has been tremendous technological progress in computer vision and in the computational power and network bandwidth of and Smartphone platforms. The synergy of these advances stands to revolutionize the way people find information and interact with the physical world. However, these technologies arenot yet fully in the hands of the visually impaired, arguably the population that could benefit the most from these developments. Part of the barrier to progress in this area has been that computer vision can accurately handle only a small fraction of thetypical images coming from a cell phone camera. To cope with these limitations and make any-image recognition possible, IQ Engines will develop a hybrid system that uses both computer vision and crowdsourcing: if the computer algorithms are not able to understand an image, then the image is sent to a unique crowdsourcing network of people for image analysis. The proposed research includes specific aims to both develop advanced computer vision algorithms for object recognition and advanced crowdsourced networks optimized to the needs of the visually impaired community. This approach combines the speed and accuracy of computer vision with the robustness and understanding of human vision, ultimately providing the user fast and accurate information about the content of any image. PUBLIC HEALTH RELEVANCE: The image recognition technology developed in this application will enable the visually impaired to access visual information using a mobile phone camera, a device most people already have. Transforming the camera into an intelligent visual sensor will lower the cost of assistance and improve the quality of life of the visually impaired community through increased independence and capabilities.