Honeywell International Inc.

Petr Cisar

Kirk A Schloegel

Yunqian Ma

Jiri Rojicek

Valerie Guralnik

Petr Stluka

Roland Miezianko

Kirk Schloegel

A system for information discovery of items, such as individuals or objects, from video-based tracks. The system may compute similarities of characteristics of the items and present the results in a matrix form. A similarity portrayal may have nodes representing the items with edges between the nodes. The edges may have weights in the form of vectors indicating similarities of the characteristics between the nodes situated at the ends of the edges. The edges may be augmented with temporal and spatial properties from the tracks which cover the items. These properties may play a part in a multi-objective presentation of information about the items in terms of a negative or supportive basis. The presentation may be partitioned into clusters which may lead to a merger of items or tracks. The system may pave a way for higher-level information discovery such as video-based social networks.

System for information discovery in video-based data

A system for social network construction. Video analytics and association may be used to develop a social network. Also, social groups may be developed from temporal proximity of persons. In the case of several social networks, they may be collapsed into one network or a weighted graph that mining algorithms can handle.

Social network construction based on data association

Methods (), algorithms, and implementing systems () for guided annotation of process deviations by an individual responsible for annotations (e.g. process operator) for a manufacturing process having an associated computing system () including a historical database (stored in memory). The historical database is provided () including data associated with a plurality of process performance measures including stored deviations for the plurality of process measures, stored quantitative deviation attributes thereof and at least one cause associated with each of the stored deviations process performance measures. The method includes the individual responsible for annotation selecting at least one present deviation (D) to be annotated responsive to a deviation in a target process performance measure (T) selected from the plurality of process measures (). Automatically, the stored quantitative attributes associated with D, T, and a target value or target interval for T are retrieved (). The database is searched to identify a plurality of similar deviations from the stored deviations determined to be similar to D and the plurality of similar deviations are statistically processed to generate a listing indicating a relative likelihood of the at least one cause associated with the plurality of similar deviations (). The listing is presented to the individual to prompt selection of at least one present cause for D from the cause(s) associated with the plurality of similar deviations ().

Guided entry system for individuals for annotating process deviations

Growing community of inventors

Suchdol nad Luznici, Czechia

Petr Cisar