System and method for expressive language and developmental disorder assessment

• US 8,938,390 B2
• Filed: 02/27/2009
• Issued: 01/20/2015
• Est. Priority Date: 01/23/2007
• Status: Active Grant

• Current Assignee: LENA Foundation
• Original Assignee: LENA Foundation
• Inventors: Xu, Dongxin D., Paul, Terrance D.
• Primary Examiner(s): Desir, Pierre-Louis
• Assistant Examiner(s): KOVACEK, DAVID M
• Application Number:
  US12/395,567
• Publication Number:
  US 20090208913A1
• Time in Patent Office: 2,153 Days
• Field of Search: 704 1- 10, 704231-235, 704236-239, 704/240, 704243-245, 704250-255, 704270-271, 704/276, 704E17001-E17016, 704E11001-E11007, 434/156, 434/185
• US Class Current: 704/245
• CPC Class Codes: A61B 5/168 Evaluating attention deficit, hyperactivity
• CPC Class Codes: A61B 5/4803 Speech analysis specially adapted for diagnostic purposes
• CPC Class Codes: A61B 5/7264 Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems neural networks per se G06N3/00; expert systems per se G06N5/00
• CPC Class Codes: A61B 5/7267 involving training the classification device
• CPC Class Codes: G10L 17/26 Recognition of special voice characteristics, e.g. for use in lie detectors; Recognition of animal voices
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems

Abstract

In one embodiment, a method for detecting autism in a natural language environment using a microphone, sound recorder, and a computer programmed with software for the specialized purpose of processing recordings captured by the microphone and sound recorder combination, the computer programmed to execute the method, includes segmenting an audio signal captured by the microphone and sound recorder combination using the computer programmed for the specialized purpose into a plurality recording segments. The method further includes determining which of the plurality of recording segments correspond to a key child. The method further includes determining which of the plurality of recording segments that correspond to the key child are classified as key child recordings. Additionally, the method includes extracting phone-based features of the key child recordings; comparing the phone-based features of the key child recordings to known phone-based features for children; and determining a likelihood of autism based on the comparing.

Claims

1. A method for detecting autism in a natural language environment using a microphone, sound recorder, and a computer programmed with software for the specialized purpose of processing recordings captured by the microphone and sound recorder combination, the computer programmed to execute a method comprising:

(a) segmenting an audio signal captured by the microphone and sound recorder combination using the computer programmed for the specialized purpose into a plurality of recording segments;

(b) determining which of the plurality of recording segments correspond to a key child;

(c) determining which of the plurality of recording segments that correspond to the key child are classified as key child recordings;

(d) extracting phone-based features of the key child recordings;

(e) comparing the phone-based features of the key child recordings to known phone-based features for children, the phone-based features corresponding to predetermined clusters of child speech resembling phones wherein the predetermined clusters of child speech resembling phones are clustered according to an unsupervised clustering method;

(f) determining a likelihood of autism based on the comparing of (e);

(g) extracting acoustic parameters of the key child recordings; and

(h) comparing the acoustic parameters of the key child recordings to known acoustic parameters for children, wherein the determining of (f) also is based on the comparing of (h).

2. The method of claim 1 wherein the acoustic parameters include canonical syllables, and the acoustic parameters are divided into four categories, including a rhythm/ syllabicity category, a low spectral tilt and high pitch control category, a wide formant bandwidth and low pitch control category, and a duration of islands within utterances category.

3. The method of claim 2 wherein the rhythm/syllabicity category includes a voiced parameter, a canonical syllable transitions category, and a spectral entropy category.

4. The method of claim 1 wherein the comparing includes Logical Regression Analysis.

5. The method of claim 1 wherein the comparing includes Linear Discriminate Analysis.

6. The method of claim 1 wherein the key child recordings are broken down into cry, vegetative sound, and vocalization categories;

and the extracting of (d) is applied only to the vocalization category of the key child recordings.

7. The method of claim 1, further comprising:

(i) performing phone analysis on the key child recordings.

8. The method of claim 1, further comprising:

(i) transforming a display of a user to display the likelihood of autism.

9. The method of claim 1, further comprising:

(j) transforming an information storage device to store the likelihood of autism.

10. The method of claim 1 wherein the phone-based features are represented by a plurality of feature vectors.

11. The method of claim 10 wherein the comparing of (e) includes the plurality of feature vectors compared to the known phone-based features for children to return a plurality of results, wherein there is a result of the plurality of results for each of the plurality of feature vectors;

and the plurality of results are averaged for use in the determining of (f).

12. The method of claim 10 wherein the plurality of feature vectors are averaged to create a single feature vector for use in the comparing of (e).

13. The method of claim 1 wherein there are 64 clusters of predetermined clusters of child speech for children above 15 months and 16 clusters of predetermined clusters of child speech for children under 15 months.

14. The method of claim 1 wherein unsupervised means that the clustering is based on the statistical characteristics of data gathered for determining the predetermined clusters of child speech.

15. The method of claim 1 wherein unsupervised means that a professional does not classify the predetermined clusters of child speech as phones.

16. A system for detecting autism in a natural language environment, the system comprising:

a microphone configured to capture a sound signal of a key child to create a plurality of audio signals;

a sound recorder configured to store the plurality of audio signals; and

a computer programmed with software for the specialized purpose of processing recordings captured by the microphone and sound recorder combination, the computer programmed to execute a method including;

(a) segmenting audio signals captured by the microphone and sound recorder combination using the computer programmed for the specialized purpose into a plurality of recording segments;

(b) determining which of the plurality of recording segments correspond to the key child;

(c) determining which of the plurality of recording segments that correspond to the key child are classified as key child recordings;

(d) extracting phone-based and transparent features of the key child recordings;

(e) comparing the phone-based and transparent features of the key child recordings to known phone-based and cluster-based transparent features for children, wherein the cluster-based transparent features are developed according to cluster-based transparent parameter analysis; and

(f) determining a likelihood of autism based on the comparing of (e); and

a display of a user, configured to display the likelihood of autism.

17. The system of claim 16 wherein the phone-based features are represented by a plurality of feature vectors.

18. The method of claim 17 wherein the plurality of feature vectors are averaged to create a single feature vector for use in the comparing of (e).

19. The system of claim 16 wherein the comparing of (e) includes the plurality of feature vectors compared to the known phone-based features for children to return a plurality of results, wherein there is a result of the plurality of results for each of the plurality of feature vectors;

and the plurality of results are averaged for use in the determining of (f).

20. The system of claim 16 wherein the method further includes:

(g) extracting acoustic parameters of the key child recordings; and

(h) comparing the acoustic parameters of the key child recordings to known acoustic parameters for children, wherein the determining of (f) also is based on the comparing of (h).

21. The system of claim 16 wherein there are 64 clusters of predetermined clusters of child speech for children above 15 months and 16 clusters of predetermined clusters of child speech for children under 15 months.

22. A method for detecting autism in a natural language environment using a microphone, sound recorder, and a computer programmed with software for the specialized purpose of processing recordings captured by the microphone and sound recorder combination, the computer programmed to execute a method comprising:

(a) segmenting an audio signal captured by the microphone and sound recorder combination using the computer programmed for the specialized purpose into a plurality of recording segments;

(b) determining which of the plurality of recording segments correspond to a key child;

(c) determining which of the plurality of recording segments that correspond to the key child are classified as key child recordings;

(d) extracting phone-based features of the key child recordings, the phone-based features corresponding to predetermined clusters of child speech resembling phones, wherein the predetermined clusters of child speech resembling phones are clustered according to an unsupervised clustering method;

(e) comparing the phone-based features of the key child recordings to known phone-based features for children; and

(f) determining a likelihood of autism based on the comparing of (e).