Human face detection adopting small scale convolutional netural network module in embedded system and head posture angle assessment

Abstract

The invention provides embodiments of a human face detection and head posture angle assessment combined system adopting a small scale hardware convolutional netural network (CNN) module in an embeddedsystem. The small scale hardware CNN module is, for example, an embedded CNN module in the Hi3519 chip of Hess Semiconductor. In some embodiments, the face detection and head posture angle assessmentcombined system disclosed in the present application is used to jointly perform multiple tasks: detecting most or all faces in a sequence of video frames, generating a posture angle assessment for the detected face, tracking the detected person's face of the same person in the sequence of video frames and generating a "best posture" assessment of the tracked person. The face detection and postureassessment joint system can be implemented in resource-limited embedded systems, such as smart camera systems that only integrate one or more small-scale CNN modules. The system proposed by the present application and the sub-image-based technology combine to implement a small-scale, low-cost CNN module to perform multiple face detection and face recognition tasks on high-resolution input images.

Description

[0001] Priority claims and related patent applications [0002] This patent application is U.S. patent application 15 / 657, 109 under examination (application name: Face Detection Using Small-scale Convolutional Neural Network (CNN) Modules for Embedded Systems); Filing Date: July 21, 2017) and claims priority under 35 U.S.C. 120 to the pending U.S. patent application 15 / 657,109. The pending U.S. patent application 15 / 657,109, pursuant to 35 U.S.C.119(e), claims U.S. provisional patent application 62 / 428,497 (Application Title: Convolutional Neural Networks Based on Finite-Resolution Small-Scale CNN Modules ( CNN) (CONVOLUTIONAL NEURAL NETWORKS (CNN) BASED ON RESOLUTION-LIMITED SMALL-SCALE CNN MODULES); filing date: November 30, 2016). The contents of all the above patent applications are incorporated into this application by reference and made a part of this application. [0003] This application is also related to pending US patent application 15 / 441,194 (Application Title: ...

AI Technical Summary

Problems solved by technology

Note that each stage of bounding box correction within a cascaded CNN requires additional computational overhead

Furthermore, in this cascaded CNN, the inherent correlation of face detection and face alignment is ignored

Images