18339408. MEMORY ORIENTED GAUSSIAN PROCESS BASED MULTI-OBJECT TRACKING (QUALCOMM Incorporated)
MEMORY ORIENTED GAUSSIAN PROCESS BASED MULTI-OBJECT TRACKING
Organization Name
Inventor(s)
Rajeev Yasarla of San Diego CA (US)
Varun Ravi Kumar of San Diego CA (US)
Senthil Kumar Yogamani of Headford (IE)
MEMORY ORIENTED GAUSSIAN PROCESS BASED MULTI-OBJECT TRACKING
This abstract first appeared for US patent application 18339408 titled 'MEMORY ORIENTED GAUSSIAN PROCESS BASED MULTI-OBJECT TRACKING
Original Abstract Submitted
An apparatus for multi-object tracking determines a current representation of a current object in a current image. The apparatus computes a joint Gaussian distribution between the current representation of the current object and a previous representation stored in one or more memory buffers, wherein the previous representation was determined from a previous image. The apparatus updates the one or more memory buffers based on the joint Gaussian distribution. For example, the apparatus determines whether to remove or replace the previous representation in the one or more memory buffers based on values of a covariance matrix of the joint Gaussian distribution.
