Method and apparatus for providing sign information

• US 9,111,142 B2
• Filed: 05/21/2013
• Issued: 08/18/2015
• Est. Priority Date: 09/27/2012
• Status: Expired due to Fees

• Current Assignee: Electronics and Telecommunications Research Institute
• Original Assignee: Electronics and Telecommunications Research Institute
• Inventors: Yoon, Ho Sub, Kim, Jae Hong, Chung, Myung Ae, Lee, Dong Jin
• Primary Examiner(s): Desire, Gregory M
• Application Number:
  US13/899,512
• Publication Number:
  US 20140086491A1
• Time in Patent Office: 819 Days
• Field of Search: 382/190, 382/103
• US Class Current: 1/1
• CPC Class Codes: G06F 2218/08 Feature extraction
• CPC Class Codes: G06V 20/63 Scene text, e.g. street names
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/168 Smoothing or thinning of the pattern; Skeletonisation
• CPC Class Codes: G06V 30/1823 using vector-coding
• CPC Class Codes: G10L 13/00 Speech synthesis; Text to speech systems

Abstract

A method and apparatus for proving sign information are disclosed. The sign information providing method includes: extracting a first sign from an input image, wherein the first sign is pre-defined; extracting a second sign representing information corresponding to the first sign around the location of the first sign, from the input image; and providing at least one piece of information of information about the first sign and information about the second sign in the form of voice. Accordingly, a user may correctly recognize information expressed by a sign.

Claims

1. A sign information providing method comprising:

extracting a first sign from a first area within an input image of an object, wherein the first sign is pre-defined;

extracting a second sign from a second area of the input image, the second sign representing information corresponding to the first sign around a location of the first sign on the object; and

providing at least one piece of information about the first sign and information about the second sign in the form of a voice message.

2. The sign information providing method of claim 1, wherein the extracting of the first sign comprises:

extracting a first area at which the first sign is located;

removing a noise signal included in the first area;

extracting feature information from the first area from which the noise signal has been removed, and recognizing a plurality of first sign candidates based on the feature information; and

extracting the first sign satisfying an authentication algorithm from among the plurality of first sign candidates.

3. The sign information providing method of claim 2, wherein the extracting of the first sign satisfying the authentication algorithm comprises extracting the first sign satisfying at least one of a Support Vector Machine (SVM)-based authentication algorithm and a Hidden Markov Model (HMM)-based authentication algorithm.

4. The sign information providing method of claim 1, wherein the extracting of the second sign from the input image comprises:

extracting a plurality of second area candidates at which the second sign is located, around the first area at which the first sign is located;

extracting a second area from among the plurality of second area candidates based on at least one piece of information of brightness information and chroma information of the plurality of second area candidates; and

extracting the second sign located in the second area.

5. The sign information providing method of claim 4, wherein the extracting of the second sign located in the second area comprises:

removing a noise signal included in the second area;

extracting feature information from the second area from which the noise signal has been removed, and recognizing a plurality of second sign candidates based on the feature information; and

extracting the second sign satisfying an authentication algorithm from among the plurality of second sign candidates.

6. The sign information providing method of claim 5, wherein the extracting of the second sign satisfying the authentication algorithm comprises extracting the second sign satisfying at least one of a Support Vector Machine (SVM)-based authentication algorithm and a Hidden Markov Model (HMM)-based authentication algorithm.

7. The sign information providing method of claim 1, wherein the extracting of the first sign comprises extracting an arrow indicating the direction of an exit, as the first sign.

8. The sign information providing method of claim 1, wherein the extracting of the second sign from the input image comprises extracting a figure representing the number of an exit, as the second sign.

9. A sign information providing apparatus comprising:

an extractor configured to extract a first sign from a first area within an input image of an object, wherein the first sign is pre-defined, and to extract a second sign from a second area of the input image, the second sign representing information corresponding to a first sign around the location of the first sign on the object; and

a provider configured to provide at least one piece of information of information about the first sign and information about the second sign in the form of a voice message.

10. The sign information providing apparatus of claim 9, wherein the extractor extracts a first area at which the first sign is located, removes a noise signal included in the first area, extracts feature information from the first area from which the noise signal has been removed, recognizes a plurality of first sign candidates based on the feature information, and extracts the first sign satisfying an authentication algorithm from among the plurality of first sign candidates.

11. The sign information providing apparatus of claim 9, wherein the extractor extracts a plurality of second area candidates at which the second sign is located, around the first area at which the first sign is located, extracts a second area from among the plurality of second area candidates based on at least one piece of information of brightness information and chroma information of the plurality of second area candidates, and extracts the second sign located in the second area.

12. The sign information providing apparatus of claim 11, wherein when the extractor extracts the second sign, the extractor removes a noise signal included in the second area, extracts feature information from the second area from which the noise signal has been removed, recognizes a plurality of second sign candidates based on the feature information, and extracts the second sign satisfying an authentication algorithm from among the plurality of second sign candidates.

13. The sign information providing apparatus of claim 9, wherein the extractor extracts an arrow indicating the direction of an exit, as the first sign, and extracts a figure representing the number of the exit, as the second sign.