Abstract: This disclosure describes techniques to count objects across multiple zones of arbitrary shapes. The techniques include operations comprising: detecting an object in a first zone of a plurality of zones of an area; determining that the object has moved from the first zone to a second zone of the plurality of zones; determining that movement of the object from the first zone to the second zone fails to satisfy a zone transition criterion for updating a count value and time criterion; and in response to determining that the movement of the object from the first zone to the second zone fails to satisfy the zone transition criterion, preventing the count value from being updated.

Abstract: This disclosure describes techniques to detect an object. The techniques include operations comprising: receiving an image captured by overhead camera; identifying a region of interest (ROI) of a plurality of regions within the image; selecting an object classifier from a plurality of object classifiers based on a position of the identified ROI relative to the overhead camera; and applying the selected object classifier to the identified ROI; and detecting presence of the object within the ROI in response to applying the selected object classifier to the identified ROI.

Abstract: This disclosure describes techniques to perform convolutional neural networks (CNNs) on embedded devices. The techniques include operations comprising: accessing DNN information including definition of layers and weights of the DNN; obtaining cache or memory information for one or more cache or memory levels of the resource constrained embedded device; and configuring the DNN to be loaded onto the one or more cache or memory levels of the resource constrained embedded device based on the cache or memory information and the DNN information.

Abstract: This disclosure describes techniques to count objects across multiple zones of arbitrary shapes. The techniques include operations comprising: detecting an object in a first zone of a plurality of zones of an area; determining that the object has moved from the first zone to a second zone of the plurality of zones; determining that movement of the object from the first zone to the second zone fails to satisfy a zone transition criterion for updating a count value and time criterion; and in response to determining that the movement of the object from the first zone to the second zone fails to satisfy the zone transition criterion, preventing the count value from being updated.

Abstract: This disclosure describes techniques to detect an object. The techniques include operations comprising: receiving an image captured by overhead camera; identifying a region of interest (ROI) of a plurality of regions within the image; selecting an object classifier from a plurality of object classifiers based on a position of the identified ROI relative to the overhead camera; and applying the selected. object classifier to the identified ROI; and detecting presence of the object within the ROI in response to applying the selected object classifier to the identified ROI.

Abstract: A method includes determining a position and length of a non-zero run in a row of a pixel map. The method also includes determining a number of neighbors for the non-zero run in a preceding row, based at least in part on the position and the length. In addition, the method includes updating a correspondence map of the non-zero run and a correspondence map of a first neighbor of the non-zero run, based at least in part on a correspondence map of a second neighbor of the non-zero run, in response to a determination that the non-zero run has at least two neighbors in the preceding row.

Abstract: In one aspect, there is disclosed a digital signal processor and method performed by the same for performing object detection, including facial detection, in a reduced number of clock cycles. The method comprises using Sobel edge detection to identify regions with many edges, and classifying those regions as foreground candidates. Foreground candidates are further checked for vertical or horizontal symmetry, and symmetrical windows are classified as face candidates. Viola-Jones type facial detection is then performed only on those windows identified as face candidates.

Abstract: The Hough transform for circles can be implemented in a manner that avoids random access to the Hough accumulator array by successively identifying center candidates in each line of the image based on edge pixels in corresponding lines voting on the line of interest.

Abstract: In one aspect, there is disclosed a digital signal processor and method performed by the same for performing object detection, including facial detection, in a reduced number of clock cycles. The method comprises using Sobel edge detection to identify regions with many edges, and classifying those regions as foreground candidates. Foreground candidates are further checked for vertical or horizontal symmetry, and symmetrical windows are classified as face candidates. Viola-Jones type facial detection is then performed only on those windows identified as face candidates.

Abstract: A method includes determining a position and length of a non-zero run in a row of a pixel map. The method also includes determining a number of neighbors for the non-zero run in a preceding row, based at least in part on the position and the length. In addition, the method includes updating a correspondence map of the non-zero run and a correspondence map of a first neighbor of the non-zero run, based at least in part on a correspondence map of a second neighbor of the non-zero run, in response to a determination that the non-zero run has at least two neighbors in the preceding row.

Abstract: The Hough transform for circles can be implemented in a manner that avoids random access to the Hough accumulator array by successively identifying center candidates in each line of the image based on edge pixels in corresponding lines voting on the line of interest.