The initial release of the Appscio™ MPF has been deployed with the company's first customer.  Currently Appscio is conducting a closed beta development process with several algorithm developer partners.  The primary objective is to expand the metadata formats and features supported.

Plans

Prior to opening up the project to the general algorithm development community we expect to have ported a number of common audio and video algorithms to the framework.  After the closed beta, we plan the following:

1. OpenCV - initial components to wrap common tasks e.g., moving data to/from IplImage, optical flow, Canny and Sobel edge detection, smoothing/blurring, color space transformations, Haar object detection, Support Vector Machine classification, ... list will continue to grow, intent is to provide most of the OpenCV higher-level fuctions (see http://opencv.willowgarage.com/wiki/ )
2. Haar-based classifier cascades for human body - face detect from OpenCV, also see http://www.lienhart.de/Source_Code/source_code.html
3. PocketSphinx - continuous stream speech-to-text (see http://www.speech.cs.cmu.edu/pocketsphinx/ )
4. LSCOM Concept extraction (see http://en.wikipedia.org/wiki/Large_Scale_Concept_Ontology_for_Multimedia )
5. Video OCR - finding text image blocks in video for routing to external OCR engines (from Movie Content Analysis project, see http://www.informatik.uni-mannheim.de/pi4/projects/MoCA/downloads-textse... ) [GPL only]