This publication describes techniques and apparatuses, implemented on a digital image capture device, directed at minimizing power consumption and central processing unit (CPU) cycles during image capture and processing events. An image sensor on the device captures a scene as a frame and generates raw image data. An on-device image signal processor (ISP) receives the raw image data and generates a statistics output (“stats output”) that includes image statistics for the frame. The stats output further includes a descriptive tag for the image statistics, saved in a header of the stats output. Software implemented on the device (e.g., a Statistics Manager) receives the stats output, parses the descriptive tag from the header, compares the descriptive tag to one or more previous descriptive tags, and determines if a change in the stats output is greater than a threshold. Upon determining that the change in the stats output is less than the threshold, the Statistics Manager determines that processing of the stats output by an Image Processing Module (e.g., 3A algorithms, other ISP software algorithms) is not necessary. Upon determining that the change in the stats output is greater than the threshold, the Statistics Manager determines processing of the stats output by the Image Processing Module is necessary. Through the use of such techniques and apparatuses, an image capture device can avoid unnecessary processing of stats outputs.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Mohtasham, Hossein and Velarde, Ruben, "Real-time Image Signal Processor Stats Management to Save Power and CPU Cycles", Technical Disclosure Commons, (June 10, 2020)