Back to results
US 9,311,522 B2Granted
Apparatus and method for improving face recognition ratio
AssigneeSamsung Electronics Co., Ltd.
InventorsSeung-yun Lee, I-saac Lee
Life of this patent
Estimating the term…
Abstract
An apparatus for improving a face recognition ratio includes a digital signal processing unit which generates a first image having a size different from that of a live-view image and, when the size of a face detected from the first image is smaller than a reference size, maps information about the face detected from the first image on the live-view image and performs smile and/or blink recognition.
Claims11
An apparatus for improving a face recognition ratio, the apparatus comprising a digital signal processor that:
- generates a live-view image at a first point in time;
- resizes the live-view image to generate a first image;
- stores the first image in a first buffer;
- updates the live-view image at a second point in time;
- if a pixel area of a face detected from the first image is less than a reference pixel area, maps information about the face detected from the first image on the updated live-view image based on a ratio between a pixel area of the first image and a pixel area of the updated live-view image, and performs at least one of smile and blink recognition on the updated live-view image;
- and otherwise,
- if the pixel area of the face detected from the first image is greater than or equal to the reference pixel area,
- performs at least one of smile and blink recognition on the first image, wherein the pixel area of the first image is less than the pixel are of the live-view image.
5 more elements
An apparatus for improving a face recognition ratio, the apparatus carrying out steps comprising a digital signal processor that:
- generates a live-view image at a first point in time;
- resizes the live-view image to generate a first image,
- resizes the first image to generate a second image;
- stores the first image in a first buffer;
- stores the second image in a second buffer;
- updates the live-view image at a second point in time;
- wherein the pixel area of the first image is less than that of the live-view image and the pixel area of the second image is less than that of the live-view image;
- if a pixel area of a face detected from the second image is less than a reference pixel area, displays a face detected from the first image within the updated live-view image overlapping a face of the updated live-view image, and
- performs at least one of a smile and blink recognition on the updated live-view image;
- otherwise,
- if the pixel area of the face detected from the second image is greater than or equal to the reference pixel area,
- displays the face detected from the second image within the updated live-view image overlapping a face of the updated live-view image, and
- performs at least one of a smile and blink recognition on the updated live-view image.
10 more elements
A method for improving a face recognition ratio, the method comprising:
- generating a live-view image at a first point in time;
- resizing the live-view image to generate a first image;
- storing the first image in a first buffer;
- updating the live-view image at a second point in time;
- performing at least one of a smile and blink recognition on the updated live-view image by mapping information about a face detected from the first image on the updated live-view image based on a ratio between a pixel area of the first image and a pixel area of the updated live-view image when a pixel area of a face detected from the first image is smaller than a reference pixel area;
- and otherwise
- performing at least one of the smile and blink recognition on the first image when the pixel area of the face detected from the first image is greater than the reference pixel area,
- wherein the pixel area of the first image is less than the pixel area of the live-view image.
5 more elements
A method for improving a face recognition ratio, the method comprising:
- generating a live-view image at a first point in time
- resizing the live-view image to generate a first image, wherein the pixel area of the first image is less than that of the live-view image and the pixel area of the second image is less than that of the live-view image;
- resizing the first image to generate a second image;
- storing the first image in a first buffer;
- storing the second image in a second buffer;
- updating the live-view image at a second point in time;
- performing at least one of a smile and blink recognition on the updated live-view image by displaying a face detected from the first image within the updated live-view image overlapping a face of the updated live-view image if a pixel area of a face detected from the second image is less than a reference pixel area;
- and
- otherwise
- performing the at least one of the smile and blink recognition on the updated live-view image by displaying the face detected from the second image within the updated live-view image overlapping a face of the updated live-view image if the pixel area of the face detected from the second image is greater than or equal to the reference pixel area.
7 more elements
Dependencies are inferred from the claim text.