Summary of the invention
The purpose of the embodiment of the present invention is that providing one kind can be improved in order to solve the above-mentioned problems in the prior art
The intelligent level of dish-washing machine and the technical solution of washing effect.
To achieve the goals above, first aspect of the embodiment of the present invention provides a kind of control method of washing of dish-washing machine, institute
The method of stating includes:
Image is shot in dish-washing machine, includes article to be cleaned in the image;
Described image is calculated, to obtain quantity, type and the soiled condition of article to be cleaned in described image;
Detect the weight of the article to be cleaned;
The weight of quantity, type and soiled condition and the article to be cleaned based on the article to be cleaned calculates institute
State the washing parameter of dish-washing machine;
The dish-washing machine is controlled to run according to the washing parameter.
Preferably, the step of quantity of the calculating article to be cleaned includes:
Color segmentation is carried out to described image;
Contour detecting is carried out to the image after color segmentation;
The shape of article to be cleaned in Image outline identification described image based on detection, to determine in described image
The quantity of article to be cleaned.
Preferably, the step of type and soiled condition of the calculating article to be cleaned includes:
Identification modeling is carried out to the article to be cleaned in described image, with the model of the determination article to be cleaned;
Carry out convolution algorithm based on model of the neural network trained in advance to the article to be cleaned with obtain it is described to
The type and soiled condition of article-cleaning.
Preferably, the washing parameter of the dish-washing machine passes through the to be cleaned article of the FUZZY ALGORITHMS FOR CONTROL based on input
The weight of quantity, type and soiled condition and the article to be cleaned is calculated.
Preferably, the washing parameter includes one or more of following: cleaning hydraulic pressure, scavenging period, cleaning strength
And steaming time.
Preferably, the method also includes:
Perceive the light intensity of dish-washing machine local environment;
The fan operation parameter of the dish-washing machine is set according to the light intensity of the dish-washing machine local environment, so that described
Blower in dish-washing machine is run according to the fan operation parameter.
Second aspect of the embodiment of the present invention provides a kind of controlling arrangement in washing of dish-washing machine, the controlling arrangement in washing packet
It includes:
Photographing module includes article to be cleaned in the image for shooting image in dish-washing machine;
Computing module, for calculating described image, to obtain quantity, the type of article to be cleaned in described image
And soiled condition;
Weight detecting module, for detecting the weight of the article to be cleaned;
Washing parameter computing module, for quantity, type and soiled condition based on the article to be cleaned and described
The weight of article to be cleaned calculates the washing parameter of the dish-washing machine;
Control module is run for controlling the dish-washing machine according to the washing parameter.
Preferably, the photographing module includes camera, and the camera is arranged in the interior sidewall surface of the dish-washing machine
The heart.
Preferably, the photographing module further includes the light compensating lamp around each camera arrangement.
Preferably, the computing module includes:
Color segmentation module, for carrying out color segmentation to described image;
Profile detection module, for carrying out contour detecting to the image after color segmentation;
Shape recognition module, for the shape of the article to be cleaned in the Image outline identification described image based on detection,
To determine the quantity of the article to be cleaned in described image.
Preferably, the computing module further include:
Identify modeling module, it is described to clear to determine for carrying out identification modeling to the article to be cleaned in described image
The model of washing product;
Convolutional neural networks, the convolutional neural networks are trained in advance, are rolled up for the model to the article to be cleaned
Operation is accumulated to obtain the type and soiled condition of the article to be cleaned.
Preferably, the washing parameter computing module is FUZZY ALGORITHMS FOR CONTROL module.
Preferably, the washing parameter includes one or more of following: cleaning hydraulic pressure, scavenging period, cleaning strength
And steaming time.
Preferably, the controlling arrangement in washing further include:
Illuminant module, for perceiving the light intensity of dish-washing machine local environment;
Fan operation parameter setting module, for washing the dishes according to the setting of the light intensity of the dish-washing machine local environment
The fan operation parameter of machine, so that the blower in the dish-washing machine is run according to the fan operation parameter.
The third aspect of the embodiment of the present invention provides a kind of dish-washing machine, including controlling arrangement in washing, the controlling arrangement in washing
For the controlling arrangement in washing of the dish-washing machine according to second aspect of the embodiment of the present invention.
Technical solution provided in an embodiment of the present invention has the following beneficial effects:
The embodiment of the present invention obtains object to be cleaned in dish-washing machine by the image comprising article to be cleaned in shooting dish-washing machine
Quantity, type and the soiled condition of product, the washing parameter of dish-washing machine is arranged in conjunction with the weight of article to be cleaned, so that washing
Bowl machine can be run according to the washing parameter, and the washing parameter of dish-washing machine is based on the quantity of article to be cleaned, type, soiled condition
It is calculated with weight, the cleaning task of the article to be cleaned under different situations can be better adapted to, so that in dish-washing machine
Article to be cleaned is cleaned cleaner, improves the cleaning effect and intelligent level of dish-washing machine.
The other feature and advantage of the embodiment of the present invention will the following detailed description will be given in the detailed implementation section.
Specific embodiment
The specific embodiment of the embodiment of the present invention is described in detail below.It should be understood that described herein
Specific embodiment be merely to illustrate and explain the present invention embodiment, be not intended to restrict the invention embodiment.
- Fig. 5 refering to fig. 1, first aspect of the embodiment of the present invention provide a kind of control method of washing of dish-washing machine, this method packet
It includes:
Image is shot in dish-washing machine, includes article to be cleaned in the image;When specifically, with dishwasher article,
First article to be cleaned is placed on the support frame in such as dish-washing machine, is then closed dish washing machine door, starting is for example pressed or turned round
Turn the start key on dishwasher control panel, starts washing procedure, after washing procedure starting, camera can for example be continuously shot and wash
It include the image of article to be cleaned in bowl machine, the image data of shooting passes through data line real-time transmission to main algorithm chip.
Described image is calculated, to obtain quantity, type and the soiled condition of article to be cleaned in described image;Tool
Body, main algorithm chip is implanted into computer vision algorithms make and model such as depth convolutional neural networks mould trained in advance
Type calculates image data by computer vision algorithms make after the image in the dish-washing machine for receiving camera shooting,
And then it can be concluded that article to be cleaned quantity, by depth convolutional neural networks model trained in advance to the image of input into
Row calculates, and then can obtain the type and soiled condition of article to be cleaned in image.
Detect the weight of the article to be cleaned;Such as weight sensor can be installed in dish-washing machine, in starting dish-washing machine
Washing procedure after, weight sensor measures the weight of article to be cleaned in dish-washing machine, and the weight data measured is transmitted
To main algorithm chip.
The weight of quantity, type and soiled condition and the article to be cleaned based on the article to be cleaned calculates institute
State the washing parameter of dish-washing machine;Prestore the algorithm model set in main algorithm chip, the input data of algorithm model be to
Quantity (projected area of the article i.e. to be cleaned in two dimensional image), the type of article to be cleaned and the soiled condition of article-cleaning
And the weight of article to be cleaned, output parameter are the washing parameter of dish-washing machine;By preset algorithm model according to defeated
The parameter entered calculates the washing parameter of dish-washing machine, and washing parameter is then sent to controller chip.
It controls the dish-washing machine to run according to the washing parameter, specifically, it is defeated that controller chip receives main algorithm chip
Washing parameter out, and control corresponding executive component in dish-washing machine and run according to the washing parameter.Washing parameter for example can be with
For cleaning strength, scavenging period, cleaning hydraulic pressure and steaming time etc., cleaning strength can for example correspond to outlet pipe upper switch valve
Aperture, aperture is bigger, and cleaning strength is higher, and scavenging period can for example correspond to the runing time of washing pump, cleans hydraulic pressure example
The running speed of washing pump can be such as corresponded to, steaming time can for example correspond to the runing time of steam generation device, generation
Steam for example can be used for the bactericidal effect of tableware to be cleaned.
The embodiment of the present invention obtains object to be cleaned in dish-washing machine by the image comprising article to be cleaned in shooting dish-washing machine
Quantity, type and the soiled condition of product, the washing parameter of dish-washing machine is arranged in conjunction with the weight of article to be cleaned, so that washing
Bowl machine can be run according to the washing parameter, and the washing parameter of dish-washing machine is based on the quantity of article to be cleaned, type, soiled condition
It is calculated with weight, the cleaning task of the article to be cleaned under different situations can be better adapted to, so that in dish-washing machine
Article to be cleaned is cleaned cleaner, improves the cleaning effect and intelligent level of dish-washing machine.
Refering to fig. 1 and Fig. 2 can calculate the quantity of the article to be cleaned in a preferred embodiment in the following way,
Main algorithm chip first carries out face to the image taken based on the two dimensional image after the two dimensional image for receiving camera shooting
Color segmentation, so that by article to be cleaned and other regions such as dish-washing machine inner wall in image, the machine support that washes the dishes is separated;Then right
Image after segmentation carries out contour detecting and is also known as edge detection;Finally in the image of the Image outline identification shooting based on detection
The shape of article to be cleaned, so that it is determined that the quantity of the article to be cleaned in image, the quantity of article to be cleaned can for example pass through
The area of the target image of extraction, that is, projected area of the article to be cleaned in two dimensional image indicates, when the perspective plane calculated
When product is bigger, illustrate that the quantity of tableware to be cleaned is more;When the profile of article to be cleaned is extracted in identification, region segmentation, that is, color
There is the irregular situation of concave or convex in the profile that the inaccuracy of segmentation normally results in extraction, in order to the recessed of profile itself
It is convex to be repaired, when identifying the image of article to be cleaned, polygon approach algorithm can be used to be fitted the side of the profile of extraction
The calculating accuracy for the quantity for treating article-cleaning further can be improved to improve the accuracy of image recognition in turn in boundary's curve.
Refering to fig. 1 and Fig. 3, in a preferred embodiment, the type of article to be cleaned and dirty can be calculated in the following way
Dirty situation, firstly, the programming of neural network is carried out using Qt compiler in linux system, with the DL frame collection Torch of open source
Or Caffe frame realizes convolutional neural networks, passes through the figure of the different degree of fouling of for example different articles to be cleaned of mass data
Piece carrys out training convolutional neural networks, obtains preferable regression result, and carries out the cutting work of convolution kernel algorithm, by cutting
RNN algorithm is implanted into main algorithm chip.When calculating the type and soiled condition of article to be cleaned, first to be cleaned in image
Article carries out identification modeling to determine the model of article to be cleaned in image, rolls up using the model of the article to be cleaned as being input to
The object module of product neural network carries out convolution algorithm by the model that convolutional neural networks trained in advance treat article-cleaning
To export the type and soiled condition of article to be cleaned.Specifically, convolutional neural networks are for example exportable uses data 0-10 institute
Type such as 1. bowls and chopsticks of the degree of fouling data of expression and article to be cleaned, 2. cups, 3. fruit, 4. knives and forks, 5. other.
Wherein, it indicates that degree of fouling is lower when exporting 0, indicates that degree of fouling is more serious when exporting 10, and export the number between 0 to 10
According to when illustrate degree of fouling between the degree of fouling shown in 0 and 10.
In one embodiment, the main algorithm chip for example can be CPU A83T ARM Cortex-A7octa-
core,512KB L1 cache 1MB L2 cache;The road GPUPowerVR SGX544MP1 Comply with OpenGL ES
2.0,OpenCL 1.x,DX 9_3.Loss function during training neural network for example can be with are as follows: MSE:The initiation parameter of convolution kernel by W and b according to~U (- sqrt (3/k),
Sqrt (3/k)) initialization, wherein k is the connection sum of W and b.
In a preferred embodiment, can the article to be cleaned by FUZZY ALGORITHMS FOR CONTROL based on input quantity, kind
The weight of class and soiled condition and the article to be cleaned calculates the washing parameter of the dish-washing machine.
Specifically, the suction parameter of the FUZZY ALGORITHMS FOR CONTROL is the degree of fouling data of convolutional neural networks model output
The category codes 1-5 of numerical value and article to be cleaned between 0-10, the weight information and pass through meter that weight sensor exports
(projected area of article to be cleaned is the projected area for the article to be cleaned that calculation machine vision algorithm is calculated in the present embodiment
The quantity of article to be cleaned);
After receiving above-mentioned data, FUZZY ALGORITHMS FOR CONTROL module is according to the perspective planes of 1. articles to be cleaned of input
The weight information that the type and soiled condition of the article to be cleaned of product, 2. convolutional neural networks feedback, 3. weight sensors return,
The fuzzy set for establishing domain U (x1, x2, x3), obtain fuzzy control rule (i.e. the washing parameter of dish-washing machine, including scavenging period,
Clean hydraulic pressure, cleaning strength, steaming time etc.), fuzzy control rule is then transferred to controller chip by serial ports, so as to
In control chip according to fuzzy control rule control dish-washing machine operation.
Wherein, the subordinating degree function selection ridge type in FUZZY ALGORITHMS FOR CONTROL is distributed osculant subordinating degree function:
In a preferred embodiment, the control method of washing of the dish-washing machine further include:
Perceive the light intensity of dish-washing machine local environment;Such as illuminant module can be set in dish-washing machine to perceive environment light
Line intensity;
The fan operation parameter of the dish-washing machine is set according to the light intensity of the dish-washing machine local environment, so that described
Blower in dish-washing machine is run according to the fan operation parameter, fan operation parameter for example may include fan operation speed, when
Between etc..In this way, can start air-dried extremely simple mode at night guarantees that night washes to reduce night fan operation speed and time
Noise is minimum when bowl machine operation.
Control method of washing based on the dish-washing machine that first aspect of the embodiment of the present invention provides, second party of the embodiment of the present invention
Face provides a kind of controlling arrangement in washing of dish-washing machine, and the controlling arrangement in washing includes:
Photographing module includes article to be cleaned in the image for shooting image in dish-washing machine;Specifically, described to take the photograph
As module for example may include camera, when with dishwasher article, first article to be cleaned is placed in such as dish-washing machine
On support frame, it is then closed dish washing machine door, starting for example presses or reverse the start key on dishwasher control panel, starting washing
Program, after washing procedure starting, camera can for example be continuously shot the image in dish-washing machine comprising article to be cleaned, the figure of shooting
As data pass through data line real-time transmission to main algorithm chip.
Computing module, for calculating described image, to obtain quantity, the type of article to be cleaned in described image
And soiled condition;Specifically, the computing module for example may include the main algorithm chip, and main algorithm chip has been implanted into calculating
Machine vision algorithm and model such as depth convolutional neural networks model trained in advance, in the dish-washing machine for receiving camera shooting
After interior image, image data is calculated by computer vision algorithms make, and then it can be concluded that article to be cleaned quantity,
The image of input is calculated by depth convolutional neural networks model trained in advance, and then can be obtained in image to clear
The type and soiled condition of washing product.
Weight detecting module, for detecting the weight of the article to be cleaned;Such as weight biography can be installed in dish-washing machine
Sensor, after the washing procedure of starting dish-washing machine, weight sensor measures the weight of article to be cleaned in dish-washing machine, and will measurement
To weight data pass to main algorithm chip.
Washing parameter computing module, for quantity, type and soiled condition based on the article to be cleaned and described
The weight of article to be cleaned calculates the washing parameter of the dish-washing machine;The algorithm model set is prestored in main algorithm chip,
The input data of algorithm model is the quantity (projected area of the article i.e. to be cleaned in two dimensional image) of article to be cleaned, to clear
The type and soiled condition of washing product and the weight of article to be cleaned, output parameter are the washing parameter of dish-washing machine;By pre-
The algorithm model first set calculates the washing parameter of dish-washing machine according to the parameter of input, and washing parameter is then sent to controller
Chip.
Control module is run for controlling the dish-washing machine according to the washing parameter, specifically, the control module example
It such as may include controller chip, controller chip receives the washing parameter of main algorithm chip output, and controls corresponding in dish-washing machine
Executive component according to the washing parameter run.Washing parameter for example can for cleaning strength, scavenging period, cleaning hydraulic pressure and
Steaming time etc., cleaning strength can for example correspond to the aperture of outlet pipe upper switch valve, and aperture is bigger, and cleaning strength is higher, clearly
The runing time of washing pump can be corresponded to by washing the time for example, and cleaning hydraulic pressure can for example correspond to the running speed of washing pump, steam
Time can for example correspond to the runing time of steam generation device, and the steam of generation for example can be used for the sterilization of tableware to be cleaned
Effect.
The embodiment of the present invention obtains object to be cleaned in dish-washing machine by the image comprising article to be cleaned in shooting dish-washing machine
Quantity, type and the soiled condition of product, the washing parameter of dish-washing machine is arranged in conjunction with the weight of article to be cleaned, so that washing
Bowl machine can be run according to the washing parameter, and the washing parameter of dish-washing machine is based on the quantity of article to be cleaned, type, soiled condition
It is calculated with weight, the cleaning task of the article to be cleaned under different situations can be better adapted to, so that in dish-washing machine
Article to be cleaned is cleaned cleaner, improves the cleaning effect and intelligent level of dish-washing machine.
Refering to Fig. 4 and Fig. 5, it is preferable that the camera 1 is arranged at the interior sidewall surface center of the dish-washing machine, such as
It is arranged in the side wall geometric center position of the cuboid liner in dish-washing machine, its liner has in normal use with dish-washing machine
(it is generally fitted with dish washing machine door in front of liner, therefore front is without side wall) for upper and lower, left and right and five, rear side wall, it is described
The inner sidewall of dish-washing machine for example can be the left side wall or right side wall of liner, in this way can be in order to detecting the projection of article to be cleaned
Area.
In a preferred embodiment, in order to enable the image that camera 1 takes is more clear, preferably to identify and examine
Article to be cleaned in altimetric image.The photographing module further includes light compensating lamp (not shown), and supplement lamp is around camera cloth
It sets around camera, carries out light filling for shooting image for camera.Specifically, light compensating lamp can be for example high brightness
LED light can for example arrange 6 light compensating lamps around each camera.
In a preferred embodiment, the computing module includes:
Color segmentation module, for carrying out color segmentation to described image;
Profile detection module, for carrying out contour detecting to the image after color segmentation;
Shape recognition module, for the shape of the article to be cleaned in the Image outline identification described image based on detection,
To determine the quantity of the article to be cleaned in described image.
Refering to fig. 1 and Fig. 2, wherein the quantity of the article to be cleaned can for example pass through the area of the target image of extraction
Projected area of the article i.e. to be cleaned in two dimensional image indicates, when the projected area calculated is bigger, illustrates to be cleaned
The quantity of tableware is more;When the profile of article to be cleaned is extracted in identification, region segmentation, that is, color segmentation inaccuracy would generally
Cause the profile extracted the irregular situation of concave or convex occur, in order to which the bumps to profile itself are repaired, is identifying
When the image of article to be cleaned, polygon approach algorithm can be used to be fitted the boundary curve of the profile of extraction, to improve in turn
The calculating accuracy for the quantity for treating article-cleaning further can be improved in the accuracy of image recognition.
In a preferred embodiment, the computing module further include:
Identify modeling module, it is described to clear to determine for carrying out identification modeling to the article to be cleaned in described image
The model of washing product;
Convolutional neural networks, the convolutional neural networks are trained in advance, are rolled up for the model to the article to be cleaned
Operation is accumulated to obtain the type and soiled condition of the article to be cleaned.
Specifically the programming of neural network can be carried out using Qt compiler in linux system, is used with Fig. 3 refering to fig. 1
DL frame collection Torch or the Caffe frame of open source realizes convolutional neural networks, passes through for example different articles to be cleaned of mass data
The pictures of different degree of fouling carry out training convolutional neural networks, obtain preferable regression result, and carry out convolution kernel algorithm
Work is cut, the RNN algorithm of cutting is implanted into main algorithm chip.It is first when calculating the type and soiled condition of article to be cleaned
Identification modeling first is carried out to determine the model of article to be cleaned in image, by the article to be cleaned to the article to be cleaned in image
Model as the object module for being input to convolutional neural networks, treat article-cleaning by convolutional neural networks trained in advance
Model carry out convolution algorithm to export the type and soiled condition of article to be cleaned.Specifically, convolutional neural networks are for example
The exportable degree of fouling data represented by data 0-10 and the type of article to be cleaned such as 1. bowls and chopsticks, 2. cups, 3.
Fruit, 4. knives and forks, 5. other.Wherein, it indicates that degree of fouling is lower when exporting 0, indicates that degree of fouling is more serious when exporting 10,
And illustrate degree of fouling between the degree of fouling shown in 0 and 10 when exporting the data between 0 to 10.
In one embodiment, the computing module for example can be CPU A83T ARM Cortex-A7octa-
core,512KB L1 cache 1MB L2 cache;The road GPUPowerVR SGX544MP1 Comply with OpenGL ES
2.0,OpenCL 1.x,DX 9_3.Loss function during training neural network for example can be with are as follows: MSE:The initiation parameter of convolution kernel by W and b according to~U (- sqrt (3/k),
Sqrt (3/k)) initialization, wherein k is the connection sum of W and b.
In a preferred embodiment, the washing parameter computing module for example can be FUZZY ALGORITHMS FOR CONTROL module, described
The suction parameter of FUZZY ALGORITHMS FOR CONTROL be convolutional neural networks model output degree of fouling data 0-10 between numerical value and
The category codes 1-5 of article to be cleaned, the weight information and calculated by computer vision algorithms make that weight sensor exports
To article to be cleaned projected area (in the present embodiment the projected area of article to be cleaned be article to be cleaned quantity);
After receiving above-mentioned data, FUZZY ALGORITHMS FOR CONTROL module is according to the perspective planes of 1. articles to be cleaned of input
The weight information that the type and soiled condition of the article to be cleaned of product, 2. convolutional neural networks feedback, 3. weight sensors return,
The fuzzy set for establishing domain U (x1, x2, x3), obtain fuzzy control rule (i.e. the washing parameter of dish-washing machine, including scavenging period,
Clean hydraulic pressure, cleaning strength, steaming time etc.), fuzzy control rule is then transferred to controller chip by serial ports, so as to
In control chip according to fuzzy control rule control dish-washing machine operation.
Wherein, the subordinating degree function selection ridge type in FUZZY ALGORITHMS FOR CONTROL is distributed osculant subordinating degree function:
In a preferred embodiment, the controlling arrangement in washing further include: illuminant module, for perceiving ring locating for dish-washing machine
The light intensity in border;
Fan operation parameter setting module, for washing the dishes according to the setting of the light intensity of the dish-washing machine local environment
The fan operation parameter of machine, so that the blower in the dish-washing machine is run according to the fan operation parameter.Fan operation parameter example
It such as may include fan operation speed, time.Extremely simple mode is air-dried in this way, can start at night, to reduce night blower
The speed of service and time guarantee that noise is minimum when the operation of night dish-washing machine.
Based on the controlling arrangement in washing that second aspect of the embodiment of the present invention provides, the third aspect of the embodiment of the present invention provides one
Kind dish-washing machine, which includes controlling arrangement in washing, wherein the controlling arrangement in washing is according to the embodiment of the present invention second
The controlling arrangement in washing for the dish-washing machine that aspect provides.
Dish-washing machine provided in an embodiment of the present invention can reduce the inaccuracy of people's subjectivity setting washing mode, use calculating
Machine vision independently simplifies the mode that operation reaches artificial intelligence, and Quan Zizhu, the full-automatic computer for realizing convolutional neural networks are deep
Degree study, so that dish-washing machine realizes artificial intelligence truly.
Next step machine learning mode is carried out according to computer vision and other sensors module such as weight sensor, in advance
Scavenging period, mode and the intensity for the needs measured all are adaptive variable quantities, reduce energy consumption, mention high-detergency, more supernumerary segment
Energy reaches cleaning mesh, realizes the full-automatic artificial intelligence dish-washing machine in just sincere justice.
The optional embodiment of the embodiment of the present invention is described in detail in conjunction with attached drawing above, still, the embodiment of the present invention is simultaneously
The detail being not limited in above embodiment can be to of the invention real in the range of the technology design of the embodiment of the present invention
The technical solution for applying example carries out a variety of simple variants, these simple variants belong to the protection scope of the embodiment of the present invention.
It is further to note that specific technical features described in the above specific embodiments, in not lance
In the case where shield, it can be combined in any appropriate way.In order to avoid unnecessary repetition, the embodiment of the present invention pair
No further explanation will be given for various combinations of possible ways.
It will be appreciated by those skilled in the art that implementing the method for the above embodiments is that can pass through
Program is completed to instruct relevant hardware, which is stored in a storage medium, including some instructions are used so that single
Piece machine, chip or processor (processor) execute all or part of the steps of each embodiment the method for the application.And it is preceding
The storage medium stated includes: USB flash disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory
The various media that can store program code such as (RAM, Random Access Memory), magnetic or disk.
In addition, any combination can also be carried out between a variety of different embodiments of the embodiment of the present invention, as long as it is not
The thought of the embodiment of the present invention is violated, equally should be considered as disclosure of that of the embodiment of the present invention.