发明名称 |
System and method for detection of high-interest events in video data |
摘要 |
A method for event identification in video data includes identifying a feature vector having data corresponding to at least one of a position and a direction of movement of an object in video data, generating an estimated feature vector corresponding to the feature vector using a dictionary including a plurality of basis vectors, identifying an error between the estimated feature vector and the feature vector, identifying a high-interest event in the video data in response to the identified error exceeding a threshold, and displaying the video data including the high-interest event on a video output device only in response to the error exceeding the threshold. |
申请公布号 |
US9589190(B2) |
申请公布日期 |
2017.03.07 |
申请号 |
US201213724389 |
申请日期 |
2012.12.21 |
申请人 |
Robert Bosch GmbH |
发明人 |
Ramakrishnan Naveen;Naim Iftekhar |
分类号 |
H04N7/18;H04N9/47;G06K9/00;G06K9/62 |
主分类号 |
H04N7/18 |
代理机构 |
Maginot, Moore & Beck LLP |
代理人 |
Maginot, Moore & Beck LLP |
主权项 |
1. A method for monitoring video data comprising:
identifying a feature vector of an event having data corresponding to at least one of a position and a direction of movement of an object in video data; generating an estimated feature vector corresponding to the feature vector using a dictionary that includes a plurality of basis vectors, the generating of the estimated feature vector further comprising:
performing a penalized optimization process with the identified feature vector and the plurality of basis vectors in the dictionary to generate a sparse weight vector that corresponds to the identified feature vector, the sparse weight vector including a plurality of elements with each element corresponding to a basis vector in the dictionary; andgenerating the estimated feature vector from a weighted sum of a plurality of basis vectors in the dictionary that correspond to elements in the sparse weight vector with non-zero weight values; identifying an error between the estimated feature vector and the identified feature vector; identifying a high-interest event in the video data in response to the identified error exceeding a threshold; displaying the video data that includes the high-interest event on a video output device only in response to the identified error exceeding the threshold; receiving a first signal from the video output device indicating the displayed video data do not include a high-interest event; and updating the dictionary in response to receiving the first signal, the updating further comprising:
generating a modified sparse weight vector based on the sparse weight vector to set any values that are less than a predetermined threshold from the sparse weight vector to zero;generating another estimated feature vector from another weighted sum of the plurality of basis vectors in the dictionary that correspond to elements in the modified sparse weight vector with non-zero weight values; andgenerating an additional basis vector based on a difference between the feature vector of the event and the other estimated feature vector; andstoring the additional basis vector in the dictionary. |
地址 |
Stuttgart DE |