embracing the challenge of transforming constantly expanding data into knowledge
Reallaer’s Video Training Catalog indexes and displays a rich library of video, making it easy to create organized sets of training material. The system contains tools for searching, reviewing, and exporting our extensive database of recordings alongside behavior and sensor metadata. Search categories are diverse, and can be composed to create customizable queries. A built-in video player alongside behavior bookmarks allows for speedy searching and perusal of video and various angles of view. As the user searches, videos can be flagged for export to create a special video set for user’s needs. All relevant metadata will follow the export, creating a master metadata sheet.
The Video Training Catalog has seen use in exporting data for career training processes, creating platforms for evaluating career progression, and continues to be maintained when navigating our in-house data collection of terabytes of video data.
Reallaer creates effective approaches to organizing data. Our metadata tagging support optimizes data organization solutions based on business operations. We apply systems engineering principles to structure and label data abstractions to assure accuracy in recorded data. We’ve found this to be of high value when supporting machine learning applications in providing labeled training data as well as powering independent verification and validation studies. We apply much of the same segmenting when training our own classifiers, having created our own digital asset management systems tailored for supporting machine learning. Additionally we have made strides in leveraging semantic web practices to provide a bridge to query engines and use a common language across disparate data.