Author: claimsbot

What is claimed is:

1. An image reading device comprising:
an illuminating section which emits visible light and infrared light and illuminates an original;
an imaging section which images one of light transmitted through the original and light reflected by the original;
an image sensor which divides an image imaged by the imaging section into a plurality of pixels and reads the image and outputs the image as image data;
a moving section which moves at least one of at least one portion of the imaging section, the image sensor, and the original, in an optical axis direction of the imaging section; and
a control section which, at each of a time of reading the image by the visible light and a time of reading the image by the infrared light, controls the moving section such that focus control is carried out by which an imaging position by the imaging section and a reading position of the image sensor coincide.
2. An image reading device according to claim 1, wherein on the basis of image data obtained by reading the image by the infrared light, the control section detects a position of at least one of scratch and foreign matter on the original, and on the basis of results of detection, corrects image data obtained by reading the image by the visible light.
3. An image reading device according to claim 2, wherein before correction of the image data obtained by reading the image by the visible light, the control section carries out at least one of magnification chromatic aberration correction and distortion aberration correction on the image data obtained by reading the image by the infrared light.
4. An image reading device according to claim 2, wherein before correction of the image data obtained by reading the image by the visible light, the control section detects an image positional offset amount between the image data obtained by reading the image by infrared light and the image data obtained by reading the image by the visible light, and, on the basis of the positional offset amount, corrects one of the image data obtained by reading the image by the infrared light or the image data obtained by reading the image by the visible light such that the positional offset amount becomes minimum.
5. An image reading device according to claim 3, wherein before correction of the image data obtained by reading the image by the visible light, the control section detects an image positional offset amount between the image data obtained by reading the image by infrared light and the image data obtained by reading the image by the visible light, and, on the basis of the positional offset amount, corrects one of the image data obtained by reading the image by the infrared light or the image data obtained by reading the image by the visible light such that the positional offset amount becomes minimum.
6. An image reading device according to claim 4, wherein the control section one of
detects the positional offset amount in advance, and each time the image is read, corrects, on the basis of the positional offset amount, one of the image data obtained by reading the image by the infrared light and the image data obtained by reading the image by the visible light such that the positional offset amount becomes minimum, and
each time the image is read, detects the positional offset amount, and corrects, on the basis of the positional offset amount, one of the image data obtained by reading the image by the infrared light and the image data obtained by reading the image by the visible light such that the positional offset amount becomes minimum.
7. An image reading device according to claim 1, wherein the control section acquires in advance a focus position for a time of image reading by the visible light and a focus position for a time of image reading by the infrared light, by controlling the illuminating section and the moving section such that focus control in a case using each the visible light and the infrared light is carried out, and
controls the moving section such that, at each time reading the image recorded on the original by the respective visible light and infrared light, at least one of at least one portion of the imaging section, the image sensor and the original moves to each position which is based on the respective focus positions acquired in advance.
8. An image reading device according to claim 1, wherein the control section acquires in advance a focus position for a time of image reading by one of the visible light and the infrared light, by controlling the illuminating section and the moving section such that focus control in a case using the one of the visible light and the infrared light is carried out, and
controls the moving section such that, at a time of reading the image by the one of the visible light and the infrared light, at least one of at least one portion of the imaging section, the image sensor and the original moves to a position which is based on the focus position acquired in advance, and controls the moving section such that, at a time of reading the image by the another of the visible light and the infrared light, at least one of at least one portion of the imaging section, the image sensor and the original moves to a position which is offset, by a predetermined offset amount which is based on a design value of the imaging section, from the position which is based on the focus position acquired in advance.
9. An image reading device according to claim 1, wherein the at least one portion of the imaging section is, in a case in which the imaging section is formed to include a single focal point lens, the single focal point lens, or is, in a case in which the imaging section is formed to include a zoom lens, at least one portion of the zoom lens.
10. An image reading device according to claim 1, wherein the imaging section is provided with a transparent parallel plate which can change the imaging position by the imaging section by being inserted onto and withdrawn from a position on an optical axis of the imaging section, and
the moving section inserts the transparent parallel plate onto and withdraws the transparent parallel plate from the position on the optical axis of the imaging section.
11. An image reading device according to claim 1, wherein the illuminating section one of
illuminates the original by selectively emitting the visible light and the infrared light, and
illuminates the original by simultaneously emitting the visible light and the infrared light.
12. An image reading device comprising:
an illuminating section which emits visible light and infrared light and illuminates an original;
an imaging section which images one of light transmitted through the original and light reflected by the original, the imaging section being provided with a transparent parallel plate which can change an imaging position by being inserted onto and withdrawn from a position on an optical axis of the imaging section;
an image sensor which divides an image imaged by the imaging section into a plurality of pixels and reads the image and outputs the image as image data;
a moving section which inserts the transparent parallel plate onto and withdraws the transparent parallel plate from the position on the optical axis of the imaging section; and
a control section which, at each of a time of reading the image by the visible light and a time of reading the image by the infrared light, controls the moving section such that focus control is carried out by which the imaging position by the imaging section and a reading position of the image sensor coincide.
13. An image reading method which illuminates visible light and infrared light onto an original, and reads an image recorded on the original on the basis of one of light transmitted through the original and light reflected by the original, the image reading method comprising the step of:
at each of a time of reading the image by the visible light and a time of reading the image by the infrared light, effecting control to move at least one of at least one portion of an imaging section which images one of the light transmitted through the original and the light reflected by the original, an image sensor which divides an image imaged by the imaging section into a plurality of pixels and reads the image and outputs the image as image data, and the original, in an optical axis direction of the imaging section, such that focus control is carried out by which an imaging position by the imaging section and a reading position of the image sensor coincide.
14. An image reading method according to claim 13, wherein on the basis of image data obtained by reading the image by the infrared light, a position of at least one of scratch and foreign matter on the original is detected, and on the basis of results of detection, image data obtained by reading the image by the visible light is corrected.
15. An image reading method according to claim 14, wherein before correction of the image data obtained by reading the image by the visible light, at least one of magnification chromatic aberration correction and distortion aberration correction is carried out on the image data obtained by reading the image by the infrared light.
16. An image reading method according to claim 14, wherein before correction of the image data obtained by reading the image by the visible light, an image positional offset amount between the image data obtained by reading the image by infrared light and the image data obtained by reading the image by the visible light, is detected, and, on the basis of the positional offset amount, one of the image data obtained by reading the image by the infrared light or the image data obtained by reading the image by the visible light is corrected such that the positional offset amount becomes minimum.
17. An image reading method according to claim 15, wherein before correction of the image data obtained by reading the image by the visible light, an image positional offset amount between the image data obtained by reading the image by infrared light and the image data obtained by reading the image by the visible light, is detected, and, on the basis of the positional offset amount, one of the image data obtained by reading the image by the infrared light or the image data obtained by reading the image by the visible light is corrected such that the positional offset amount becomes minimum.
18. An image reading method according to claim 16, wherein one of
detecting the positional offset amount in advance, and each time the image is read, correcting, on the basis of the positional offset amount, one of the image data obtained by reading the image by the infrared light and the image data obtained by reading the image by the visible light such that the positional offset amount becomes minimum, and
each time the image is read, detecting the positional offset amount, and correcting, on the basis of the positional offset amount, one of the image data obtained by reading the image by the infrared light and the image data obtained by reading the image by the visible light such that the positional offset amount becomes minimum, is performed.
19. An image reading method according to claim 13, wherein a focus position for a time of image reading by the visible light and a focus position for a time of image reading by the infrared light are acquired in advance, by controlling illuminating of the visible light and the infrared light and moving of at least one of at least one portion of the imaging section, the image sensor and the original such that focus control in a case using each the visible light and the infrared light is carried out, and
at each time reading the image recorded on the original by the respective visible light and infrared light, at least one of at least one portion of the imaging section, the image sensor and the original is controlled to move to each position which is based on the respective focus positions acquired in advance.
20. An image reading method according to claim 13, wherein a focus position for a time of image reading by one of the visible light and the infrared light is acquired in advance, by controlling illuminating of the visible light and the infrared light and moving of at least one of at least one portion of the imaging section, the image sensor and the original such that focus control in a case using the one of the visible light and the infrared light is carried out, and
at a time of reading the image by the one of the visible light and the infrared light, at least one of at least one portion of the imaging section, the image sensor and the original is controlled to move to a position which is based on the focus position acquired in advance, and at a time of reading the image by the another of the visible light and the infrared light, at least one of at least one portion of the imaging section, the image sensor and the original is controlled to move to a position which is offset, by a predetermined offset amount which is based on a design value of the imaging section, from the position which is based on the focus position acquired in advance.
21. An image reading method according to claim 13, wherein the at least one portion of the imaging section is, in a case in which the imaging section is formed to include a single focal point lens, the single focal point lens, or is, in a case in which the imaging section is formed to include a zoom lens, at least one portion of the zoom lens.
22. An image reading method according to claim 13, wherein the imaging section is provided with a transparent parallel plate which can change the imaging position by the imaging section by being inserted onto and withdrawn from a position on an optical axis of the imaging section, and
inserting the transparent parallel plate onto and withdrawing the transparent parallel plate from the position on the optical axis of the imaging section is controlled.
23. An image reading method according to claim 13, wherein the visible light and the infrared light are illuminated to the original by one of selectively emitting and simultaneously emitting.
24. An image reading method which illuminates visible light and infrared light onto an original, and reads an image recorded on the original on the basis of one of light transmitted through the original and light reflected by the original, the image reading method comprising the ‘step of:
at each of a time of reading the image by the visible light and a time of reading the image by the infrared light, effecting control to move a transparent parallel plate which is provided at an imaging section imaging one of the light transmitted through the original and the light reflected by the original on an image sensor which divides an image imaged by the imaging section into a plurality of pixels and reads the image and outputs the image as image data and which can change an imaging position by being inserted onto and withdrawn from a position on an optical axis of the imaging section, such that focus control is carried out by which the imaging position by the imaging section and a reading position of the image sensor coincide.

The claims below are in addition to those above.
All refrences to claim(s) which appear below refer to the numbering after this setence.

What is claimed:

1. A method for spotting words in a speech signal, the method comprising the steps of:
generating a first recognition score based on the speech signal and a lexicon entry for a first word, the first recognition score tracking an absolute likelihood that the first word is in the speech signal;
estimating a first background score based on the first recognition score; and
calculating a first confidence score based on a matching ratio between a first minimum recognition value and the first background score, the first confidence score tracking a noise-corrected likelihood that the first word is in the speech signal.
2. The method of claim 1 further including the step of averaging the first recognition score over a predetermined period of time.
3. The method of claim 1 further including the steps of:
dividing the first minimum recognition value by an average value for the first recognition score over a predetermined period of time such that the matching ratio results, the average value defining the first background score; and
normalizing the matching ratio;
said normalized matching ratio defining the first confidence score.
4. The method of claim 3 further including the step of locating a minimum value within the first recognition score.
5. The method of claim 4 further including the step of searching a predetermined range of the first recognition score for the minimum value such that local minimums are excluded from the matching ratio calculation.
6. The method of claim 1 further including the step of comparing the first confidence score to a predetermined confidence threshold, the first word being in the speech signal when the first confidence score exceeds the predetermined confidence threshold.
7. The method of claim 6 further including the step of spotting a second word in the speech signal.
8. The method of claim 7 further including the steps of:
generating a second recognition score based on the speech signal and a lexicon entry for a second word, the second recognition score tracking an absolute likelihood that the second word is in the speech signal;
estimating a second background score based on the second recognition score; and
calculating a second confidence score based on a matching ratio between a second minimum recognition value and the second background score, the second confidence score tracking a noise-corrected likelihood that the second word is in the speech signal.
9. The method of claim 8 further including the step of comparing the second confidence score to the predetermined confidence threshold, the second word being in the speech signal when the second confidence score exceeds the predetermined confidence threshold.
10. The method of claim 9 further including the steps of:
determining whether the first word and the second word correspond to a common time period within the speech signal; and
selecting between the first word and the second word based on the first confidence score and the second confidence score when the first word and the second word correspond to the common time period.
11. The method of claim 1 further including the step of calculating the confidence score on a frame-by-frame basis.
12. A method for calculating a word spotting confidence score for a given word, the method comprising the steps of:
dividing a minimum value of a speech recognition score by an average value of the speech recognition score over a predetermined period of time such that a matching ratio results, the average value defining an estimated background score; and
normalizing the matching ratio;
said normalized matching ratio defining the confidence score.
13. The method of claim 12 further including the step of locating the minimum value within the speech recognition score.
14. The method of claim 13 further including the step of searching a predetermined range of the recognition score for the minimum value such that local minimums are excluded from the matching ratio calculation.
15. A word spotting system comprising:
a speech recognizer for generating recognition scores based on a speech signal and lexicon entries for a plurality of words, the recognition scores tracking absolute likelihoods that the words are in the speech signal; and
a spotting module for estimating background scores based on the recognition scores;
said spotting module calculating confidence scores on a frame-by-frame basis based on matching ratios between minimum recognition values and the background scores, the confidence scores tracking noise-corrected likelihoods that the words are in the speech signal.
16. The word spotting system of claim 15 wherein the spotting module includes:
a confidence module for dividing the minimum recognition values by average values for the recognition scores such that the matching ratios result, the average values defining the background scores;
said confidence module normalizing the matching ratios such that the normalized matching ratios define the confidence scores.

We claim:

1. An automatic method for generating summaries for text documents, comprising steps of:
generating a set of sentences for a set of documents by document discourse analysis and a set of words by morphologic process;
initializing a score for each word in the set of words and for each sentence in the set of sentences;
computing the score for each word in the set of words according to the score of sentences containing it and the correlation degree between the word and the user information;
computing the score for each sentence in the set of sentences according to the score of words composing it and the position of the sentence in a section and a paragraph; and
if the sum of scores of the words and the sum of scores of the sentences change apparently, go back to the step of computing the word score; otherwise continuing;
outputting the top-ranked sentences as the summary of the set of documents, the top-ranked words as the keywords list of the set of documents.
2. An automatic method according to claim 1, wherein the step of computing the score for each word comprises:
computing the score for each word in the set of words according to the linguistic salience of the word to the user profile.
3. An automatic method according to claim 1, wherein the step of computing the score for each word comprises:
computing the score for each word in the set of words according to the similarity among the word, the query and topic provided by a user.
4. An automatic method according to claim 1, wherein the step of computing the score for each word comprises:
computing the score for each word in the set of words according to the similarity among the word and the terms in the titles of the documents.
5. An automatic method according to claim 1, wherein the step of computing the score for each word comprises:
computing the score for each word in the set of words according to the ratio of its occurrence number in the document to its occurrence number in the set of documents.
6. An automatic method according to claim 1, wherein the step of computing the score for each word comprises:
computing the score for each word in the set of words according to the ratio of the number of the documents including the word to the number of documents in the set of documents.
7. An automatic method according to claim 1, wherein the step of computing the score for each word comprises:
computing the score for each word in the set of words according to the weighted-average of at least two of:
the linguistic salience of the word to the user profile;
the similarity among the word, the query and topic provided by a user;
the similarity among the word and the terms in the titles of the documents;
the ratio of its occurrence number in the document to its occurrence number in the set of documents; and
the ratio of the number of the documents comprising the word to the number of documents in the set of documents.
8. A computer program product for automatically generating summaries for text documents, said computer program product comprising a computer usable medium having computer readable program code thereon, said computer readable program code comprising:
computer program code means for generating a set of sentences for a set of documents by document discourse analysis and a set of words by morphologic process;
computer program code means for initializing a score for each word in the set of words, and each sentence in the set of sentences;
computer program code means for computing the score for each word in the set of words according to the score of sentences containing it and the correlation degree between the word and the user information;
computer program code means for computing the score for each sentence in the set of sentences according to the score of words composing it and the position of the sentence in a section and a paragraph;
computer program code means for determining if the sum of scores of the words and the sum of scores of the sentences exhibit an apparent change; and
computer program code means for outputting the top-ranked sentences as the summary of the set of documents, the top-ranked words as the keywords list of the set of documents.
9. A computer program product for automatically generating summaries according to claim 8, wherein the computer program code means for computing the score for each word comprises:
computer program code means for computing the score for each word in the set of words according to the linguistic salience of the word to the user profile.
10. A computer program product for automatically generating summaries according to claim 8, wherein the computer program code means for computing the score for each word comprises:
computer program code means for computing the score for each word in the set of words according to the similarity among the word, the query and topic provided by a user.
11. A computer program product for automatically generating summaries according to claim 8, wherein the computer program code means for computing the score for each word comprises:
computer program code means for computing the score for each word in the set of words according to the similarity among the word and the terms in the titles of the documents.
12. A computer program product for automatically generating summaries according to claim 8, wherein the computer program code means for computing the score for each word comprises:
computer program code means for computing the score for each word in the set of words according to the ratio of its occurrence number in the document to its occurrence number in the set of documents.
13. A computer program product for automatically generating summaries according to claim 8, wherein the computer program code means for computing the score for each word comprises:
computer program code means for computing the score for each word in the set of words according to the ratio of the number of the documents comprising the word to the number of documents in the set of documents.
14. A computer program product for automatically generating summaries according to claim 8, wherein the computer program code means for computing the score for each word comprises:
computer program code means for computing the score for each word in the set of words according to the weighted-average of at least two of:
the linguistic salience of the word to the user profile,
the similarity among the word, the query and topic provided by a user,
the similarity among the word and the terms in the titles of the documents,
the ratio of its occurrence number in the document to its occurrence number in the set of documents, and
the ratio of the number of the documents including the word to the number of documents in the set of documents.

The claims below are in addition to those above.
All refrences to claim(s) which appear below refer to the numbering after this setence.

1. A power supply circuit for a PCI-E slot comprising:
a control chip, wherein the control chip determines the status of a motherboard, outputting a control signal;
a first electronic switch including a first terminal, a second terminal, and a third terminal, wherein the first terminal is connected to the control chip to receive the control signal and connected to a 3.3V dual power supply of the motherboard through a first resistor, the second terminal is grounded, the third terminal is connected to the 3.3V dual power supply through a second resistor; and
a second electronic switch including a first terminal, a second terminal, and a third terminal, wherein the third terminal of the first electronic switch is further connected to the first terminal of the second electronic switch, the second terminal of the second electronic switch is connected to the 3.3V dual power supply, the third terminal of the second electronic switch is connected to a PCI-E slot.
2. The power supply circuit for a PCI-E slot of claim 1, wherein the first electronic switch is a field effect transistor, the gate of the field effect transistor corresponds to the first terminal of the first electronic switch, the source of the field effect transistor corresponds to the second terminal of the first electronic switch, the drain of the field effect transistor corresponds to the third terminal of the first electronic switch.
3. The power supply circuit for a PCI-E slot of claim 1, wherein the second electronic switch is a field effect transistor, the gate of the field effect transistor corresponds to the first terminal of the second electronic switch, the source of the field effect transistor corresponds to the second terminal of the second electronic switch, the drain of the field effect transistor corresponds to the third terminal of the second electronic switch.
4. The power supply circuit for a PCI-E slot of claim 1, wherein the control chip is a super IO chip.