CN109300458A - A kind of electronics music leaf turner of automatic page turning - Google Patents
A kind of electronics music leaf turner of automatic page turning Download PDFInfo
- Publication number
- CN109300458A CN109300458A CN201811376863.3A CN201811376863A CN109300458A CN 109300458 A CN109300458 A CN 109300458A CN 201811376863 A CN201811376863 A CN 201811376863A CN 109300458 A CN109300458 A CN 109300458A
- Authority
- CN
- China
- Prior art keywords
- swimming
- section
- distance
- line
- symbol
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10G—REPRESENTATION OF MUSIC; RECORDING MUSIC IN NOTATION FORM; ACCESSORIES FOR MUSIC OR MUSICAL INSTRUMENTS NOT OTHERWISE PROVIDED FOR, e.g. SUPPORTS
- G10G7/00—Other auxiliary devices or accessories, e.g. conductors' batons or separate holders for resin or strings
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B42—BOOKBINDING; ALBUMS; FILES; SPECIAL PRINTED MATTER
- B42D—BOOKS; BOOK COVERS; LOOSE LEAVES; PRINTED MATTER CHARACTERISED BY IDENTIFICATION OR SECURITY FEATURES; PRINTED MATTER OF SPECIAL FORMAT OR STYLE NOT OTHERWISE PROVIDED FOR; DEVICES FOR USE THEREWITH AND NOT OTHERWISE PROVIDED FOR; MOVABLE-STRIP WRITING OR READING APPARATUS
- B42D9/00—Bookmarkers; Spot indicators; Devices for holding books open; Leaf turners
- B42D9/04—Leaf turners
Abstract
The present invention relates to a kind of electronics music leaf turners of automatic page turning characterized by comprising display module, the display module are connected with electronic display, and electronic display is for showing one page music score;Storage unit, the storage unit connect the CPU element for storing electronic music information, the storage unit;Power supply and power management module, the power supply can be used to power to the electronics music leaf turner, and the power management module is for managing in-line power;CPU element connects each unit of the electronics music leaf turner, and the information and control signal for each unit to the electronics music leaf turner are handled;Music score symbol extraction module, the module extract the text and note of electronic music.
Description
Technical field
The invention belongs to electronic music page turning technical fields, and in particular to a kind of electronics music leaf turner of automatic page turning.
Background technique
During the training of musical performance and performance, music score is essential tool.Player needs to play on one side,
Music score is seen on one side, and as music score is constantly stirred in the progress of performance.
The generation of electronics music leaf turner can preferably solve the problems, such as this, but most of music score are all paperys at present,
It will be on the music leaf turner of electronic display screen using very inconvenient.And existing music leaf turner can not facilitate realization to turn over spectrum, it is unfavorable
It trains or performs in playing.
This is in place of the deficiencies in the prior art.Therefore, in view of the above-mentioned drawbacks in the prior art, provide and design one kind automatically
The electronics music leaf turner of page turning;To solve drawbacks described above in the prior art, it is necessary.
Summary of the invention
It is an object of the present invention in view of the above-mentioned drawbacks of the prior art, provide the electricity for designing a kind of automatic page turning
Sub- music leaf turner, to solve the above technical problems.
To achieve the above object, the present invention provides following technical scheme:
A kind of electronics music leaf turner of automatic page turning characterized by comprising
Display module, the display module are connected with electronic display, and electronic display is for showing one page music score;
Storage unit, for the storage unit for storing electronic music information, it is mono- that the storage unit connects the CPU
Member;
Power supply and power management module, the power supply can be used to power to the electronics music leaf turner, the power management mould
Block is for managing in-line power;
CPU element connects each unit of the electronics music leaf turner, the information for each unit to the electronics music leaf turner
It is handled with control signal;
Music score symbol extraction module, the module extract the text and note of electronic music, specifically includes the following steps:
S1: spectral line deletes step;Specifically includes the following steps:
S1.1: spectral line detection is obtained with line parameters,
The Run- Length Coding of vertical direction is carried out to image (retouching line direction is vertical direction, is scanned by column from left to right);
The frequency that black (prospect) distance of swimming changes with length is counted, histogram, the corresponding distance of swimming of peak-peak in figure are obtained
Length is line width;
The frequency that white (background) distance of swimming changes with length is counted, histogram, the corresponding distance of swimming of peak-peak in figure are obtained
Length is separation of spectra;
S1.2: carrying out spectral line deletion based on the analysis of figure section topological relation,
Define 1: the adjacent distance of swimming, if at least one pixel is individually present in two distances of swimming in adjacent two scan lines in left and right
Row coordinate having the same, then claiming the two distances of swimming that there is syntople.
Define 2: run-length adjacency graph, if G represents run-length adjacency graph, it is binary group < V, E a >, i.e. G=< V, E >.Its
In: V is node set, and node is distance of swimming object;E is line set, records the syntople between the distance of swimming.
Define 3: abutting one by one, if two distances of swimming with syntople no longer with its in the scan line where them
His distance of swimming has syntople, and the distance of swimming has syntople one by one.
Define 4: figure section is referred to as figure section by the set that one group of adjacent distance of swimming forms, and these distances of swimming meet the following conditions:
Continuous arrangement, the adjacent distance of swimming in left and right have syntople one by one to each distance of swimming from left to right, while other than the both ends distance of swimming and set
The distance of swimming does not have syntople one by one.
Definition 5: Parent Submap section, if the last one distance of swimming of figure section A and first distance of swimming of figure section B have adjacent close
System, then claiming A is father's figure section of B, and B is the subgraph section of A.
Define 6: figure section adjacent map is binary group < V, E a >, i.e. G=< V, E > if G represents figure section adjacent map.Wherein:
V is node set, and node is figure section object;E is line set, records the set membership between neighbor map section.
Definition 7: figure section topological classification, if figure section, without Parent Submap section, the topological classification of the figure section is stand alone;If figure
There are father's figure section or subgraph sections for section, and number is respectively less than 2, then the topological classification of the figure section is referred to as branching type, is otherwise intersection point type.
The algorithm for analyzing progress spectral line deletion based on figure section topological relation is as follows:
It vertically scans, Run- Length Coding is carried out to image, establishes run-length adjacency graph.
On the basis of run-length adjacency graph, distance of swimming polymerization is realized with connection distance of swimming growth algorithm;This is a depth-first
Process, foundation figure section simultaneously generates figure section adjacent map, records the topological classification of each figure section.
When figure section meets following four condition simultaneously, it is determined as spectrum section, the pixel number of figure section is deleted from image
According to.
Condition 1: line width constraint;Figure section is averaged distance of swimming thickness less than certain threshold value (2 times of line widths)
Condition 2: morphological constraints;The thickness that figure section surrounds frame is less than certain threshold value (2.5 times of line widths)
Condition 3: position constraint;The difference of horizontal center line and position of spectral line that figure section surrounds frame is less than certain threshold value (1 times of spectrum
Line line width).
Condition 4: topological classification;Constraints graph segment type cannot be the intersection point type defined in 7.
S2: dry segmentation step is accorded with;
It vertically scans, to musical score image, (unless otherwise specified, the initial pictures hereafter in algorithm refer both to spectral line
Image after removing) vertical runs coding is carried out, if length is more than certain threshold value (2 times of separations of spectra) in the black pixel distance of swimming, mark
Remember that the distance of swimming is candidate's Fu Gan distance of swimming;
It scans in the horizontal direction, horizontal runs coding is carried out to image, retrieve in the candidate Fu Gan distance of swimming where pixel
Horizontal runs, thus to obtain the profile point of pixel right boundary;
If candidate's Fu Gan distance of swimming profile point meets the following conditions, it is determined that the candidate Fu Gan distance of swimming is that symbol is dry.
Condition 5: line width constraint puts the distance between profile point of two sides less than certain threshold value (2 times of line widths).
Condition 6: symmetric constraints, profile point is to the center line bilateral symmetry relative to the candidate Fu Gan distance of swimming.
Condition 7: consistency constraint, while the profile point for meeting condition 5,6 must be continuous and reaches (1.5 times of certain length
Line width).
S3: symbol head segmentation step;
It scans in the horizontal direction, horizontal runs coding is carried out to image.It deletes length and is less than certain threshold value (between 3 times of spectral lines
Away from) the black pixel distance of swimming of level, it is therefore an objective to remove connector head and accord with the symbol stem portion of beam, realize symbol head and accord with the separation of beam.
Connected domain mark is carried out to image, for each connected domain, if width is greater than certain threshold value (between 3 times of spectral lines
Away from), and any one edge line then deletes the connected domain close to straight line up and down.The purpose of the step is that removal has obvious symbol
The connected domain of beam feature avoids Fu Liang from accidentally being known after the cutting of next step for symbol head.
It is done with symbol as segmentation lines, all connected domains of cutting, the symbol head of lateral adhesion is separated after execution.
It is greater than the connected domain of certain threshold value (2 times of separations of spectra) using line between spectral line or spectral line as segmentation lines cutting height, holds
The symbol head of longitudinal adhesion is separated after row.
Whether has symbol head feature (symbol head connected domain by the connected domain after three length, height and area Indexs measure cuttings
Long is about that 1.5 times of separations of spectra, a height of 1 times of separation of spectra, elemental area and connected domain surround frame area ratio as 0.8).
S4: symbol beam segmentation step;
It scans in the horizontal direction, horizontal runs coding is carried out to image.It deletes length and is less than certain threshold value (between 3 times of spectral lines
Away from) the black pixel distance of swimming of level, it is therefore an objective to remove connector head and accord with the symbol stem portion of beam, realize symbol head and accord with the separation of beam.
By thickness, length, edge line three Indexs measure images of straightness in connected domain whether have Fu Liangte
(connected domain thickness connects not less than 2 times of spectral line thickness, length not less than in separation of spectra, symbol beam lower edges line at least one sign
Nearly straight line).
The Run- Length Coding of vertical direction is carried out to the symbol beam blocks image split;
The frequency that the distance of swimming changes with length in statistics Run- Length Coding, generates frequency histogram, chooses maximum peak in histogram
It is worth corresponding run length as single symbol cantilever thickness.
S5: word lines Graph Extraction step, specifically includes the following steps:
S5.1: Word Input step specifically includes:
Region division is carried out to music score according to position of spectral line, rectangular area made of five spectral line boundaries are surrounded is denoted as spectrum
The area Xian Zu;Region far from spectral line group and spectral line group two sides, is denoted as header area, and the region between spectral line group and spectral line group is denoted as
Literal field;
Connected domain mark is carried out to image, the connected domain in header area is directly placed into text set;Surround frame and spectral line group
It is non-legible that area has overlapping connected domain to be considered as, and does not deal with;It is located at the connected domain in literal field for surrounding frame, executes following step
It is rapid:
Frame is surrounded in screening has the connected domain of character size feature (length in 0.5~5 times of separation of spectra, highly 1~3
Times separation of spectra), character set is added;
Connected domain trifling in text is merged into character, for short and small connected domain, if it surrounds frame and text set
In certain surround frame and there are overlapping or two to surround less than 0.3 times of the horizontally or vertically gap separations of spectra between frames, then merge encirclement
Frame;
Text strings merger, if two adjacent encirclement frame vertical direction have overlapping in text set, and horizontal clearance is less than
0.3 times of separation of spectra, then merge encirclement frame.
S5.2: the step of note symbol extraction;It specifically includes:
Note symbol is identified using the geometrical characteristic of connected domain in image;
Bar includes tie, crescendo diminuendo mark and omits mark line in music score, has obviously geometry
Feature: angle is small close to horizontal, line width and longer dimension;Therefore, after text object is deleted, it can directly utilize and be connected in image
The geometrical characteristic in domain (surrounds width of frame not less than 3 times of separations of spectra, surrounds the average length of vertical runs in frame to be differentiated
2) less than 0.5 times separation of spectra, the ratio for surrounding frame length and height are greater than.
Page turning node audio generation module, by every page of the last one text extracted in music score symbol extraction module and most
Page turning node of the combination as this page of the latter note, and by the combination producing page turning node audio;
External audio input module receives external audio data, and is sent to CPU element;
Audio contrast module, the audio signal data that external audio input module is received and page turning node audio carry out
It compares;If identical, page turning signal is sent to CPU element and carries out page turning.
Preferably, the electronic display is electronic ink screen.
Preferably, push-button unit, is provided with multiple keys, shown including page turning of score is carried out by pressing mode
The turnover key of operation carries out the selection key of music score rotation by pressing mode and cancels key;A variety of page turn over operation are provided.
Preferably, communication module, the communication module connects the CPU element, and the communication module passes through wireless network
Network acquisition exterior content is simultaneously interactive to the CPU element by exterior content.
The beneficial effects of the present invention are, by receiving external audio data information, and by its with page turning node audio into
Row compares, and when external audio data is just page turning node audio, regards as external played to the last one node of this page;
Page-turning instruction then, which is sent, to CPU element carries out page turning.
In addition, design principle of the present invention is reliable, structure is simple, has very extensive application prospect.
It can be seen that compared with prior art, the present invention have substantive distinguishing features outstanding and it is significant ground it is progressive, implementation
Beneficial effect be also obvious.
Detailed description of the invention
Fig. 1 is a kind of control principle drawing of the electronics music leaf turner of automatic page turning provided by the invention.
Wherein, 1- display module, 2- storage unit, 3- power supply and power management module, 4-CPU unit, 5- music score symbol
Extraction module, 6- page turning node audio generation module, 7- external audio input module, 8- audio contrast module, 9- push-button unit,
10- communication module.
Specific embodiment
The present invention will be described in detail with reference to the accompanying drawing and by specific embodiment, and following embodiment is to the present invention
Explanation, and the invention is not limited to following implementation.
As shown in Figure 1, a kind of electronics music leaf turner of automatic page turning provided by the invention characterized by comprising
Display module 1, the display module are connected with electronic display, and electronic display is for showing one page music score;
Storage unit 2, the storage unit connect CPU element 4 for storing electronic music information, the storage unit;
Power supply and power management module 3, the power supply can be used to power to the electronics music leaf turner, the power management mould
Block is for managing in-line power;
CPU element 4 connects each unit of the electronics music leaf turner, the letter for each unit to the electronics music leaf turner
Breath and control signal are handled;
Music score symbol extraction module 5, the module extract the text and note of electronic music, specifically includes the following steps:
S1: spectral line deletes step;Specifically includes the following steps:
S1.1: spectral line detection is obtained with line parameters,
The Run- Length Coding of vertical direction is carried out to image (retouching line direction is vertical direction, is scanned by column from left to right);
The frequency that black (prospect) distance of swimming changes with length is counted, histogram, the corresponding distance of swimming of peak-peak in figure are obtained
Length is line width;
The frequency that white (background) distance of swimming changes with length is counted, histogram, the corresponding distance of swimming of peak-peak in figure are obtained
Length is separation of spectra;
S1.2: carrying out spectral line deletion based on the analysis of figure section topological relation,
Define 1: the adjacent distance of swimming, if at least one pixel is individually present in two distances of swimming in adjacent two scan lines in left and right
Row coordinate having the same, then claiming the two distances of swimming that there is syntople.
Define 2: run-length adjacency graph, if G represents run-length adjacency graph, it is binary group < V, E a >, i.e. G=< V, E >.Its
In: V is node set, and node is distance of swimming object;E is line set, records the syntople between the distance of swimming.
Define 3: abutting one by one, if two distances of swimming with syntople no longer with its in the scan line where them
His distance of swimming has syntople, and the distance of swimming has syntople one by one.
Define 4: figure section is referred to as figure section by the set that one group of adjacent distance of swimming forms, and these distances of swimming meet the following conditions:
Continuous arrangement, the adjacent distance of swimming in left and right have syntople one by one to each distance of swimming from left to right, while other than the both ends distance of swimming and set
The distance of swimming does not have syntople one by one.
Definition 5: Parent Submap section, if the last one distance of swimming of figure section A and first distance of swimming of figure section B have adjacent close
System, then claiming A is father's figure section of B, and B is the subgraph section of A.
Define 6: figure section adjacent map is binary group < V, E a >, i.e. G=< V, E > if G represents figure section adjacent map.Wherein:
V is node set, and node is figure section object;E is line set, records the set membership between neighbor map section.
Definition 7: figure section topological classification, if figure section, without Parent Submap section, the topological classification of the figure section is stand alone;If figure
There are father's figure section or subgraph sections for section, and number is respectively less than 2, then the topological classification of the figure section is referred to as branching type, is otherwise intersection point type.
The algorithm for analyzing progress spectral line deletion based on figure section topological relation is as follows:
It vertically scans, Run- Length Coding is carried out to image, establishes run-length adjacency graph.
On the basis of run-length adjacency graph, distance of swimming polymerization is realized with connection distance of swimming growth algorithm;This is a depth-first
Process, foundation figure section simultaneously generates figure section adjacent map, records the topological classification of each figure section.
When figure section meets following four condition simultaneously, it is determined as spectrum section, the pixel number of figure section is deleted from image
According to.
Condition 1: line width constraint;Figure section is averaged distance of swimming thickness less than certain threshold value (2 times of line widths)
Condition 2: morphological constraints;The thickness that figure section surrounds frame is less than certain threshold value (2.5 times of line widths)
Condition 3: position constraint;The difference of horizontal center line and position of spectral line that figure section surrounds frame is less than certain threshold value (1 times of spectrum
Line line width).
Condition 4: topological classification;Constraints graph segment type cannot be the intersection point type defined in 7.
S2: dry segmentation step is accorded with;
It vertically scans, to musical score image, (unless otherwise specified, the initial pictures hereafter in algorithm refer both to spectral line
Image after removing) vertical runs coding is carried out, if length is more than certain threshold value (2 times of separations of spectra) in the black pixel distance of swimming, mark
Remember that the distance of swimming is candidate's Fu Gan distance of swimming;
It scans in the horizontal direction, horizontal runs coding is carried out to image, retrieve in the candidate Fu Gan distance of swimming where pixel
Horizontal runs, thus to obtain the profile point of pixel right boundary;
If candidate's Fu Gan distance of swimming profile point meets the following conditions, it is determined that the candidate Fu Gan distance of swimming is that symbol is dry.
Condition 5: line width constraint puts the distance between profile point of two sides less than certain threshold value (2 times of line widths).
Condition 6: symmetric constraints, profile point is to the center line bilateral symmetry relative to the candidate Fu Gan distance of swimming.
Condition 7: consistency constraint, while the profile point for meeting condition 5,6 must be continuous and reaches (1.5 times of certain length
Line width).
S3: symbol head segmentation step;
It scans in the horizontal direction, horizontal runs coding is carried out to image.It deletes length and is less than certain threshold value (between 3 times of spectral lines
Away from) the black pixel distance of swimming of level, it is therefore an objective to remove connector head and accord with the symbol stem portion of beam, realize symbol head and accord with the separation of beam.
Connected domain mark is carried out to image, for each connected domain, if width is greater than certain threshold value (between 3 times of spectral lines
Away from), and any one edge line then deletes the connected domain close to straight line up and down.The purpose of the step is that removal has obvious symbol
The connected domain of beam feature avoids Fu Liang from accidentally being known after the cutting of next step for symbol head.
It is done with symbol as segmentation lines, all connected domains of cutting, the symbol head of lateral adhesion is separated after execution.
It is greater than the connected domain of certain threshold value (2 times of separations of spectra) using line between spectral line or spectral line as segmentation lines cutting height, holds
The symbol head of longitudinal adhesion is separated after row.
Whether has symbol head feature (symbol head connected domain by the connected domain after three length, height and area Indexs measure cuttings
Long is about that 1.5 times of separations of spectra, a height of 1 times of separation of spectra, elemental area and connected domain surround frame area ratio as 0.8).
S4: symbol beam segmentation step;
It scans in the horizontal direction, horizontal runs coding is carried out to image.It deletes length and is less than certain threshold value (between 3 times of spectral lines
Away from) the black pixel distance of swimming of level, it is therefore an objective to remove connector head and accord with the symbol stem portion of beam, realize symbol head and accord with the separation of beam.
By thickness, length, edge line three Indexs measure images of straightness in connected domain whether have Fu Liangte
(connected domain thickness connects not less than 2 times of spectral line thickness, length not less than in separation of spectra, symbol beam lower edges line at least one sign
Nearly straight line).
The Run- Length Coding of vertical direction is carried out to the symbol beam blocks image split;
The frequency that the distance of swimming changes with length in statistics Run- Length Coding, generates frequency histogram, chooses maximum peak in histogram
It is worth corresponding run length as single symbol cantilever thickness.
S5: word lines Graph Extraction step, specifically includes the following steps:
S5.1: Word Input step specifically includes:
Region division is carried out to music score according to position of spectral line, rectangular area made of five spectral line boundaries are surrounded is denoted as spectrum
The area Xian Zu;Region far from spectral line group and spectral line group two sides, is denoted as header area, and the region between spectral line group and spectral line group is denoted as
Literal field;
Connected domain mark is carried out to image, the connected domain in header area is directly placed into text set;Surround frame and spectral line group
It is non-legible that area has overlapping connected domain to be considered as, and does not deal with;It is located at the connected domain in literal field for surrounding frame, executes following step
It is rapid:
Frame is surrounded in screening has the connected domain of character size feature (length in 0.5~5 times of separation of spectra, highly 1~3
Times separation of spectra), character set is added;
Connected domain trifling in text is merged into character, for short and small connected domain, if it surrounds frame and text set
In certain surround frame and there are overlapping or two to surround less than 0.3 times of the horizontally or vertically gap separations of spectra between frames, then merge encirclement
Frame;
Text strings merger, if two adjacent encirclement frame vertical direction have overlapping in text set, and horizontal clearance is less than
0.3 times of separation of spectra, then merge encirclement frame.
S5.2: the step of note symbol extraction;It specifically includes:
Note symbol is identified using the geometrical characteristic of connected domain in image;
Bar includes tie, crescendo diminuendo mark and omits mark line in music score, has obviously geometry
Feature: angle is small close to horizontal, line width and longer dimension;Therefore, after text object is deleted, it can directly utilize and be connected in image
The geometrical characteristic in domain (surrounds width of frame not less than 3 times of separations of spectra, surrounds the average length of vertical runs in frame to be differentiated
2) less than 0.5 times separation of spectra, the ratio for surrounding frame length and height are greater than.
Page turning node audio generation module 6, by every page of the last one text extracted in music score symbol extraction module and
Page turning node of the combination as this page of the last one note, and by the combination producing page turning node audio;
External audio input module 7 receives external audio data, and is sent to CPU element;
Audio contrast module 8, the audio signal data that external audio input module is received and page turning node audio into
Row compares;If identical, page turning signal is sent to CPU element and carries out page turning.
In the present embodiment, the electronic display is electronic ink screen.
In the present embodiment, further includes push-button unit 9, be provided with multiple keys, find pleasure in including by pressing mode
The turnover key for composing page turning display operation carries out the selection key of music score rotation by pressing mode and cancels key;It provides more
Kind page turn over operation.
It further include communication module 10 in the present embodiment, the communication module connects the CPU element, the communication module
It is by wireless network acquisition exterior content and exterior content is interactive to the CPU element.
Disclosed above is only the preferred embodiment of the present invention, but the present invention is not limited to this, any this field
What technical staff can think does not have creative variation, and without departing from the principles of the present invention made by several improvement and
Retouching, should all be within the scope of the present invention.
Claims (4)
1. a kind of electronics music leaf turner of automatic page turning characterized by comprising
Display module, the display module are connected with electronic display, and electronic display is for showing one page music score;
Storage unit, the storage unit connect CPU element for storing electronic music information, the storage unit;
Power supply and power management module, the power supply can be used to power to the electronics music leaf turner, and the power management module is used
In management in-line power;
CPU element connects each unit of the electronics music leaf turner, the information and control for each unit to the electronics music leaf turner
Signal processed is handled;
Music score symbol extraction module, the module extract the text and note of electronic music, specifically includes the following steps:
S1: spectral line deletes step;Specifically includes the following steps:
S1.1: spectral line detection is obtained with line parameters,
The Run- Length Coding of vertical direction is carried out to image;
The frequency that the black distance of swimming changes with length is counted, histogram is obtained, the corresponding run length of peak-peak is line width;
The frequency that the white distance of swimming changes with length is counted, histogram is obtained, the corresponding run length of peak-peak is separation of spectra;
S1.2: carrying out spectral line deletion based on the analysis of figure section topological relation,
Define 1: the adjacent distance of swimming, if two distances of swimming in adjacent two scan lines in left and right are individually present at least one pixel and have
Identical row coordinate, then claiming the two distances of swimming that there is syntople;
Define 2: run-length adjacency graph, if G represents run-length adjacency graph, it is binary group < V, E a >, i.e. G=< V, E >;Wherein: V
For node set, node is distance of swimming object;E is line set, records the syntople between the distance of swimming;
Define 3: abut one by one, if with syntople two distances of swimming no longer in the scan line where them other trip
Journey has syntople, and the distance of swimming has syntople one by one;
Define 4: figure section is referred to as figure section by the set that one group of adjacent distance of swimming forms, and these distances of swimming meet the following conditions: each trip
Continuous arrangement, the adjacent distance of swimming in left and right have syntople one by one, while the distance of swimming other than the both ends distance of swimming and set to journey from left to right
Do not have syntople one by one;
Definition 5: Parent Submap section, if the last one distance of swimming of figure section A and first distance of swimming of figure section B have syntople,
Claiming A is father's figure section of B, and B is the subgraph section of A;
Define 6: figure section adjacent map is binary group < V, E a >, i.e. G=< V, E > if G represents figure section adjacent map;Wherein: V is
Node set, node are figure section object;E is line set, records the set membership between neighbor map section;
Definition 7: figure section topological classification, if figure section, without Parent Submap section, the topological classification of the figure section is stand alone;If figure section is deposited
In father's figure section or subgraph section, and number is respectively less than 2, then the topological classification of the figure section is referred to as branching type, is otherwise intersection point type;
The algorithm for analyzing progress spectral line deletion based on figure section topological relation is as follows:
It vertically scans, Run- Length Coding is carried out to image, establishes run-length adjacency graph;
On the basis of run-length adjacency graph, distance of swimming polymerization is realized with connection distance of swimming growth algorithm;Foundation figure section simultaneously generates figure section neighbour
Map interlinking records the topological classification of each figure section;
When figure section meets following four condition simultaneously, it is determined as spectrum section, the pixel data of figure section is deleted from image:
Condition 1: line width constraint;Figure section is averaged less than 2 times line widths of distance of swimming thickness;
Condition 2: morphological constraints;Less than 2.5 times line widths of thickness of figure section encirclement frame;
Condition 3: position constraint;Figure section surrounds the horizontal center line of frame and less than 1 times line width of difference of position of spectral line;
Condition 4: topological classification;Constraints graph segment type cannot be the intersection point type defined in 7;
S2: dry segmentation step is accorded with;
It vertically scans, vertical runs coding is carried out to musical score image, if length is more than between 2 times of spectral lines in the black pixel distance of swimming
Away from then marking the distance of swimming is candidate's Fu Gan distance of swimming;
It scans in the horizontal direction, horizontal runs coding is carried out to image, retrieves the water in the candidate Fu Gan distance of swimming where pixel
The flat distance of swimming, thus to obtain the profile point of pixel right boundary;
If candidate's Fu Gan distance of swimming profile point meets the following conditions, it is determined that the candidate Fu Gan distance of swimming is that symbol is dry;
Condition 5: less than 2 times line widths of the distance between profile point of two sides are put in line width constraint;
Condition 6: symmetric constraints, profile point is to the center line bilateral symmetry relative to the candidate Fu Gan distance of swimming;
Condition 7: consistency constraint, while the profile point for meeting condition 5,6 must be continuous and reaches 1.5 times of line widths;
S3: symbol head segmentation step;
It scans in the horizontal direction, horizontal runs coding is carried out to image;Delete level black pixel of the length less than 3 times of separations of spectra
The distance of swimming, removal connector head and the symbol stem portion for according with beam realize symbol head and accord with the separation of beam;
Connected domain mark is carried out to image, for each connected domain, if width is greater than 3 times of separations of spectra, and upper and lower any one
Edge line then deletes the connected domain close to straight line;
It is done with symbol as segmentation lines, all connected domains of cutting, the symbol head of lateral adhesion is separated after execution;
It is greater than the connected domain of 2 times of separations of spectra using line between spectral line or spectral line as segmentation lines cutting height, longitudinal adhesion after execution
Symbol head is separated;
Whether has symbol head feature by the connected domain after three length, height and area Indexs measure cuttings;
S4: symbol beam segmentation step;
It scans in the horizontal direction, horizontal runs coding is carried out to image;Delete level black pixel of the length less than 3 times of separations of spectra
The distance of swimming, removal connector head and the symbol stem portion for according with beam realize symbol head and accord with the separation of beam;
By thickness, length, edge line three Indexs measure images of straightness in connected domain whether have symbol beam feature;
The Run- Length Coding of vertical direction is carried out to the symbol beam blocks image split;
The frequency that the distance of swimming changes with length in statistics Run- Length Coding, generates frequency histogram, chooses peak-peak pair in histogram
The run length answered is as single symbol cantilever thickness;
S5: word lines Graph Extraction step, specifically includes the following steps:
S5.1: Word Input step specifically includes:
Region division is carried out to music score according to position of spectral line, rectangular area made of five spectral line boundaries are surrounded is denoted as spectral line group
Area;Region far from spectral line group and spectral line group two sides, is denoted as header area, and the region between spectral line group and spectral line group is denoted as text
Area;
Connected domain mark is carried out to image, the connected domain in header area is directly placed into text set;Surround frame has with spectral line group area
Overlapping connected domain is considered as non-legible, does not deal with;It is located at the connected domain in literal field for surrounding frame, executes following steps:
The connected domain that frame has character size feature is surrounded in screening, and character set is added;
Connected domain trifling in text is merged into character, for short and small connected domain, if it surrounds certain in frame and text set
Surrounding frame has overlapping or two less than 0.3 times of horizontally or vertically gap separations of spectra surrounded between frame, then merges encirclement frame;
Text strings merger, if two adjacent encirclement frame vertical direction have overlapping in text set, and horizontal clearance is less than 0.3 times
Separation of spectra then merges encirclement frame;
S5.2: the step of note symbol extraction;It specifically includes:
Note symbol is identified using the geometrical characteristic of connected domain in image;
Page turning node audio generation module, by every page of the last one text extracted in music score symbol extraction module and last
Page turning node of the combination as this page of a note, and by the combination producing page turning node audio;
External audio input module receives external audio data, and is sent to CPU element;
Audio contrast module compares audio signal data and page turning node audio that external audio input module receives
It is right;If identical, page turning signal is sent to CPU element and carries out page turning.
2. a kind of electronics music leaf turner of automatic page turning according to claim 1, which is characterized in that the electronic display
For electronic ink screen.
3. a kind of electronics music leaf turner of automatic page turning according to claim 2, which is characterized in that push-button unit is provided with
Multiple keys are carried out including the turnover key for carrying out page turning of score display operation by pressing mode by pressing mode
The selection key and cancellation key of music score rotation.
4. a kind of electronics music leaf turner of automatic page turning according to claim 3, which is characterized in that communication module, it is described logical
Believe that module connects the CPU element, the communication module is by wireless network acquisition exterior content and interactive extremely by exterior content
The CPU element.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811376863.3A CN109300458A (en) | 2018-11-19 | 2018-11-19 | A kind of electronics music leaf turner of automatic page turning |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811376863.3A CN109300458A (en) | 2018-11-19 | 2018-11-19 | A kind of electronics music leaf turner of automatic page turning |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109300458A true CN109300458A (en) | 2019-02-01 |
Family
ID=65144741
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811376863.3A Pending CN109300458A (en) | 2018-11-19 | 2018-11-19 | A kind of electronics music leaf turner of automatic page turning |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109300458A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112434177A (en) * | 2020-11-27 | 2021-03-02 | 北京邮电大学 | Three-dimensional model retrieval method and device, electronic equipment and storage medium |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20020126898A1 (en) * | 2001-01-31 | 2002-09-12 | Guo Jinhong K. | Run length based connected components and contour following for enhancing the performance of circled region extraction algorithm |
CN106384568A (en) * | 2016-11-21 | 2017-02-08 | 丹阳伦图电子技术有限公司 | Electronic music book turning device with automatic page turning function |
CN106504730A (en) * | 2016-11-19 | 2017-03-15 | 丹阳伦图电子技术有限公司 | A kind of electronics music leaf turner and its using method based on audio analysis |
-
2018
- 2018-11-19 CN CN201811376863.3A patent/CN109300458A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20020126898A1 (en) * | 2001-01-31 | 2002-09-12 | Guo Jinhong K. | Run length based connected components and contour following for enhancing the performance of circled region extraction algorithm |
CN106504730A (en) * | 2016-11-19 | 2017-03-15 | 丹阳伦图电子技术有限公司 | A kind of electronics music leaf turner and its using method based on audio analysis |
CN106384568A (en) * | 2016-11-21 | 2017-02-08 | 丹阳伦图电子技术有限公司 | Electronic music book turning device with automatic page turning function |
Non-Patent Citations (2)
Title |
---|
刘晓翔: "《乐谱图像中的音符识别方法》", 《计算机工程》 * |
刘晓翔: "《光学乐谱识别技术研究与实现》", 《中国优秀博硕士学位论文全文数据库(博士) 信息科技辑》 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112434177A (en) * | 2020-11-27 | 2021-03-02 | 北京邮电大学 | Three-dimensional model retrieval method and device, electronic equipment and storage medium |
CN112434177B (en) * | 2020-11-27 | 2023-06-20 | 北京邮电大学 | Three-dimensional model retrieval method and device, electronic equipment and storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN101251892B (en) | Method and apparatus for cutting character | |
CN102663382B (en) | Video image character recognition method based on submesh characteristic adaptive weighting | |
CN101102419B (en) | A method for caption area of positioning video | |
CN104966051B (en) | A kind of Layout Recognition method of file and picture | |
CN100474340C (en) | Image processing method and image processing device | |
CN111242027B (en) | Unsupervised learning scene feature rapid extraction method fusing semantic information | |
CN102063619B (en) | Character row extraction method and device | |
Salvatore et al. | Contour line recognition from scanned topographic maps | |
CN103577818A (en) | Method and device for recognizing image characters | |
CN110180186A (en) | A kind of topographic map conversion method and system | |
CN103034848A (en) | Identification method of form type | |
CN102930537A (en) | Image detection method and system | |
JP3411472B2 (en) | Pattern extraction device | |
CN105701490B (en) | A kind of container number adaptive location method based on image entropy | |
CN111368695A (en) | Table structure extraction method | |
CN101930544B (en) | Run adjacency table-based staff quick connected domain analysis method | |
CN109300458A (en) | A kind of electronics music leaf turner of automatic page turning | |
Lemaitre et al. | A perceptive method for handwritten text segmentation | |
CN101944180A (en) | Music note primitive segmentation method based on music note knowledge and double projection method | |
KR101793184B1 (en) | Apparatus of Fast lyric area extraction from images of printed music scores and method therefor | |
CN107392206A (en) | Rail pressed characters dividing method under a kind of working condition | |
CN101615255B (en) | Video text multi-frame interfusion method | |
Modi et al. | Text line detection and segmentation in Handwritten Gurumukhi Scripts | |
Seeri et al. | A novel approach for Kannada text extraction | |
CN109948598A (en) | Document layout intelligent analysis method and device |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination |