Self-adaptive printing method and device based on multi-modal recognition, equipment and medium

Abstract

The embodiment of the invention discloses a self-adaptive printing method and device based on multi-modal recognition, equipment and a medium. A specific embodiment of the method comprises the following steps: carrying out multi-modal identification processing on a printed article to obtain article attribute information; determining an article identifier of the printed article based on the article attribute information; self-adaptive printing parameter information corresponding to the article identifier is matched; and controlling the printing equipment to print the printed article based on the self-adaptive printing parameter information. According to the embodiment, the printing quality is improved.

Description

Technical Field

[0001] Embodiments of this disclosure relate to the field of computer technology, and more specifically to adaptive printing methods, apparatus, devices, and media based on multimodal recognition. Background Technology

[0002] In applications such as personalized printing and smart manufacturing, UV printers' adaptive printing is a technology that can automatically adjust printing parameters based on the characteristics of the object being printed. Currently, the common methods for implementing UV printing are: acquiring rough information about the item through a single type of sensor and matching it with printing parameters, or having operators manually set a uniform printing parameter scheme based on experience.

[0003] However, when using the above method to print with a UV printer, the following technical problems often arise: Recognizing objects from a single information source is inaccurate because it cannot comprehensively assess the object's material, surface characteristics, and identity information. This leads to poor compatibility of the matching printing parameters and poor print quality (for example, identifying a printed object as ordinary wood, but it is actually ABS plastic requiring special surface treatment; in this case, the printer uses parameters unsuitable for ABS, resulting in print quality problems such as ink bleeding, insufficient adhesion, and color deviation). At the same time, relying on manual experience to adjust parameters results in low printing production efficiency.

[0004] The information disclosed in this background section is only intended to enhance the understanding of the background of the inventive concept, and therefore may contain information that does not form prior art known to those skilled in the art. Summary of the Invention

[0005] The summary portion of this disclosure is intended to provide a brief overview of the concepts, which will be described in detail in the detailed description portion. This summary portion is not intended to identify key or essential features of the claimed technical solutions, nor is it intended to limit the scope of the claimed technical solutions.

[0006] Some embodiments of this disclosure propose adaptive printing methods, apparatuses, electronic devices, and computer-readable media based on multimodal recognition to address one or more of the technical problems mentioned in the background section above.

[0007] In a first aspect, some embodiments of this disclosure provide an adaptive printing method based on multimodal recognition. The method includes: performing multimodal recognition processing on the item to be printed to obtain item attribute information; determining an item identifier of the item to be printed based on the item attribute information; matching adaptive printing parameter information corresponding to the item identifier; and controlling a printing device to print the item to be printed based on the adaptive printing parameter information.

[0008] Secondly, some embodiments of this disclosure provide an adaptive printing apparatus based on multimodal recognition. The apparatus includes: an acquisition unit configured to perform multimodal recognition processing on an item to be printed to obtain item attribute information; a generation unit configured to determine an item identifier of the item to be printed based on the item attribute information; a processing unit configured to match adaptive printing parameter information corresponding to the item identifier; and an input unit configured to control a printing device to print the item to be printed based on the adaptive printing parameter information.

[0009] Thirdly, some embodiments of this disclosure provide an electronic device, including: one or more processors; and a storage device having one or more programs stored thereon, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method described in any implementation of the first aspect above.

[0010] Fourthly, some embodiments of this disclosure provide a computer-readable medium having a computer program stored thereon, wherein the program, when executed by a processor, implements the method described in any of the implementations of the first aspect above.

[0011] The above-described embodiments of this disclosure have the following beneficial effects: the adaptive printing method based on multimodal recognition in some embodiments of this disclosure improves printing quality and printing production efficiency. Specifically, the reasons for poor printing quality and low printing production efficiency are: by identifying items from a single information source, the material, surface characteristics, and identity information of the object cannot be comprehensively judged, resulting in poor accuracy of the identification results, which in turn leads to poor adaptability of the matching printing parameters and poor printing quality (e.g., ink bleeding, insufficient adhesion, color deviation, and other printing quality problems); at the same time, the printing production efficiency is low because it relies on manual experience to adjust parameters. Based on this, the adaptive printing method based on multimodal recognition in some embodiments of this disclosure first performs multimodal recognition processing on the item to be printed to obtain item attribute information. Thus, multidimensional attribute information of the item to be printed can be obtained. Next, based on the item attribute information, the item identifier of the item to be printed is determined. Thus, comprehensive analysis and judgment can be performed based on multidimensional attribute information, avoiding the identification bias caused by a single information source and improving the accuracy and reliability of item identifier (item type or identity identifier). Next, the adaptive printing parameter information corresponding to the aforementioned item identifier is matched. Thus, based on the accurately determined item identifier, the optimal set of printing parameters for that type of item is matched, i.e., adaptive printing parameter information adapted to the specific material and surface characteristics of the item being printed. Then, the printing equipment is controlled to print the item based on the aforementioned adaptive printing parameter information. Therefore, printing is performed according to the matched printing parameter information adapted to the specific material and surface characteristics of the item being printed, reducing printing defects caused by parameter mismatches and improving print quality. Furthermore, because the entire process from identifying and matching adaptive printing parameter information to executing printing is completed without manual intervention, production downtime and reliance on operator experience are significantly reduced, thus improving printing production efficiency. Attached Figure Description

[0012] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. Throughout the drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic, and elements are not necessarily drawn to scale.

[0013] Figure 1 This is a flowchart of some embodiments of the adaptive printing method based on multimodal recognition according to the present disclosure; Figure 2 This is a schematic diagram of the structure of some embodiments of the adaptive printing apparatus based on multimodal recognition according to the present disclosure; Figure 3 This is a schematic diagram of the structure of an electronic device suitable for implementing some embodiments of the present disclosure; Figure 4 These are internal product diagrams of some embodiments of the printing apparatus according to this disclosure; Figure 5 It is an internal test diagram of the printing device according to this disclosure during the printing process; Figure 6 This is an internal test printout of an adaptive printing method based on multimodal recognition, which is used in accordance with this disclosure and printed by a printing device. Detailed Implementation

[0014] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure.

[0015] It should also be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings. Unless otherwise specified, the embodiments and features described in this disclosure can be combined with each other.

[0016] It should be noted that the concepts of "first" and "second" mentioned in this disclosure are used only to distinguish different devices, modules or units, and are not used to limit the order of functions performed by these devices, modules or units or their interdependencies.

[0017] It should be noted that the terms "a" and "a plurality of" used in this disclosure are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0018] The names of messages or information exchanged between multiple devices in the embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.

[0019] This disclosure will now be described in detail with reference to the accompanying drawings and embodiments.

[0020] Figure 1 A flow 100 of some embodiments of the adaptive printing method based on multimodal recognition according to this disclosure is shown. The adaptive printing method based on multimodal recognition includes the following steps: Step 101: Perform multimodal recognition processing on the printed item to obtain item attribute information.

[0021] In some embodiments, the execution entity of the adaptive printing method based on multimodal recognition (e.g., a computing device, a central processing unit module in a UV printer) performs multimodal recognition processing on the item to be printed to obtain item attribute information. The item to be printed can be an item to be UV printed (e.g., an oak craft or an acrylic phone case).

[0022] In some optional implementations of certain embodiments, the aforementioned execution entity can perform multimodal recognition processing on the printed item through the following steps to obtain item attribute information: The first step involves collecting at least two items of object perception and recognition information from the printed item to constitute item attribute information. These at least two items of object perception and recognition information include at least two of the following: image recognition information, near-field communication (NFC) recognition information, and weight information. In practice, the executing entity can combine any two of the image recognition information, NFC recognition information, and weight information to form the item attribute information. Optionally, the executing entity can determine the image recognition information, NFC recognition information, and weight information as the item attribute information.

[0023] In some optional implementations of certain embodiments, the aforementioned execution entity may collect at least two items of object perception and recognition information of the printed item through the following steps: The first step involves acquiring a visual image of the surface of the item to be printed, and generating image recognition information based on this image. In practice, the executing entity can use an industrial camera (such as a CCD or CMOS sensor) to capture an image of the item to be printed within the printing area as the surface visual image. This image recognition information can represent surface features of the item (e.g., smooth and transparent) or identify the type of item (e.g., acrylic sheet). Then, by analyzing the texture features of pixels in the surface visual image (e.g., using a gray-level co-occurrence matrix (GLCM) to calculate texture roughness and contrast), or by analyzing the reflective / diffuse properties of the item's surface, the entity can determine whether the surface is smooth (e.g., ceramic, glass) or rough (e.g., wood, canvas), thus obtaining image recognition information. Optionally, the executing entity can compare the surface visual image to the presence of specific visual markers or feature points pre-stored in a database to obtain image recognition information. For example, some carrier pallets may have special positioning markings or markings used to identify the type of item.

[0024] The second step involves reading the NFC tag embedded in the printed item or its carrier tray to obtain near-field communication (NFC) identification information. In practice, the executing entity can use an NFC reader or RFID reader to read the NFC tag embedded in the printed item or its carrier tray to obtain NFC identification information. This NFC identification information can represent the material information of the printed item, and includes a verification identifier, such as one of "acrylic," "304 stainless steel," or "oak."

[0025] The third step is to collect the weight information of the item to be printed. In practice, the aforementioned execution entity can collect the weight information of the item to be printed by integrating a weighing sensor (such as a strain gauge sensor) into the support structure of the UV printer's printing platform. As an example, after the item is placed, the sensor converts the weight of the item into an electrical signal, which is then converted into a digital weight value by an analog-to-digital converter, serving as the weight information of the item to be printed.

[0026] In some optional implementations of certain embodiments, the aforementioned execution entity may generate image recognition information based on the aforementioned surface visual image through the following steps: The first step is to perform feature extraction processing on the aforementioned surface visual image to obtain the visual feature information of the item to be printed. In practice, the aforementioned execution entity can input the surface visual image into a pre-trained deep learning model (e.g., a convolutional neural network (CNN), such as VGG, ResNet, etc.) to obtain the visual feature information of the item to be printed. This visual feature information of the item to be printed is one or a set of structured, quantized data vectors used to represent the key physical properties (color, shape, texture, gloss) of the surface of the item to be printed.

[0027] The second step involves generating image recognition information based on the visual feature information of the printed item. In practice, the executing entity can use a pre-trained recognition model to obtain the image recognition information. This recognition model can be a classification model that takes the visual feature information of the printed item as input and the image recognition information as output.

[0028] In addressing the aforementioned vision-based adaptive selection of surface printing parameters using technical solutions, the following technical challenges arise for the application scenario: printing high-value, high-precision products or products with complex visual appearances (e.g., art reproductions, personalized luxury goods surface treatments, visually consistent coatings for precision industrial components). When acquiring images of the object's surface, inherent optical distortion of the camera lens and perspective distortion caused by the shooting angle prevent the acquired raw images from accurately reflecting the object's color distribution, texture, and gloss characteristics. Directly extracting features from distorted images leads to visual feature information deviating from the object's true physical properties. Considering the comprehensiveness and accuracy of visual feature extraction required for this application scenario, we have decided to adopt the following solution: In some optional implementations of certain embodiments, the execution entity may perform feature extraction processing on the surface visual image through the following steps to obtain visual feature information of the printed item: The first step is to perform distortion correction processing on the above surface visual image using an image distortion correction algorithm to obtain a corrected surface visual image.

[0029] The second step involves extracting color features from the aforementioned visual image of the corrected surface to obtain color feature extraction information. In practice, color feature extraction can be performed on the visual image of the corrected surface using a color histogram (such as HSV space transformation) or a convolutional neural network (CNN) to obtain color feature extraction information. This color feature extraction information can be a feature vector representing the color features of the visual image of the corrected surface.

[0030] The third step involves extracting texture features from the aforementioned visual image of the corrected surface to obtain texture feature extraction information. In practice, image texture analysis and extraction techniques based on the gray-level co-occurrence matrix (GLCM) can be used to perform co-occurrence matrix feature extraction processing on the aforementioned visual image of the corrected surface to obtain texture feature information. This texture feature information can be a feature vector representing texture features of the visual image of the corrected surface, including contrast, homogeneity, energy, entropy, and correlation.

[0031] The fourth step involves extracting gloss features from the printed object in the aforementioned corrected surface visual image to obtain gloss feature extraction information. In practice, this can be achieved by calculating the area ratio and pixel intensity distribution of the highlight region in the corrected surface visual image, and then estimating the surface gloss characteristics using a reflection model based on these factors, thus obtaining vector-based gloss characteristic information. This gloss characteristic information can refer to information describing the surface gloss level.

[0032] The fifth step involves stitching together the extracted color features, texture features, and gloss features to obtain the visual feature information of the printed item.

[0033] The above-described steps, as an inventive point of this disclosure, solve the technical problem that "when acquiring images of the surface of an object to be printed, due to the inherent optical distortion of the camera lens and the perspective distortion caused by the shooting angle, the acquired original image cannot realistically and geometrically accurately reflect the color distribution, texture structure, and gloss characteristics of the object's surface. If features are directly extracted based on the distorted image, the visual feature information of the printed object will deviate from the true physical properties of the printed object." In practice, traditional visual feature extraction schemes directly process the original image. The extracted color histogram may be distorted due to geometric distortion, texture features (such as the GLCM matrix) may be oriented incorrectly and have inaccurate statistical values ​​due to perspective distortion, and gloss feature calculations may be misjudged due to distortion of the highlight area. Ultimately, the generated visual feature vector cannot represent the true visual properties of the object's surface, leading to inappropriate selection of subsequent printing parameters. This disclosure presents a complete technical solution that starts with image geometric correction and proceeds to the accurate extraction and fusion of multimodal features of color, texture, and gloss. By establishing a geometrically accurate image benchmark from the source through an image distortion correction algorithm, color, texture, and gloss features are extracted from this benchmark and then concatenated into a comprehensive visual feature vector, ensuring that the extracted visual feature information faithfully and completely reflects the real physical properties of the object's surface.

[0034] In addressing the aforementioned technical problems by employing technical solutions, the application scenario of adaptive printing based on the material recognition of the printed object often presents the following challenges: Material recognition of the printed object's surface image yields identification information, but the acquired images typically suffer from background interference. Furthermore, the printing chamber, being a closed or semi-closed space, is prone to uneven lighting and shadows, resulting in noise in the acquired images. Additionally, the subtle differences in surface texture, gloss, and reflection between different materials make it difficult for traditional single-feature extraction methods to accurately distinguish between material types, leading to inaccurate matching of subsequent printing parameters and consequently, poor print quality. Therefore, this application scenario requires the following characteristics: suitability for material recognition of printed objects in closed or semi-closed spaces.

[0035] In some optional implementations of certain embodiments, the aforementioned execution entity may generate image recognition information based on the aforementioned surface visual image through the following steps: The first step is to perform low-light environment noise reduction processing on the aforementioned surface visual image to obtain a denoised surface visual image. In practice, the aforementioned execution entity can use Gaussian filtering image denoising technology to perform low-light environment noise reduction processing on the surface visual image to obtain a denoised surface visual image. The aforementioned denoised surface visual image can be the image obtained after performing Gaussian filtering denoising processing on the surface visual image.

[0036] The second step is to perform grayscale processing on the above-mentioned denoised surface visual image to obtain a grayscale surface visual image. The grayscale surface visual image can be the grayscale-processed denoised surface visual image.

[0037] The third step involves inputting the aforementioned grayscale surface visual image into a pre-trained instance segmentation model to obtain the information about the area of ​​the item to be printed. This instance segmentation model can be a Mask R-CNN model that takes the grayscale surface visual image as input and outputs the information about the area of ​​the item to be printed. The information about the area of ​​the item to be printed can represent the position of the item in the grayscale surface visual image, and may include various coordinates.

[0038] The fourth step is to generate a background segmentation mask based on the aforementioned information about the area of ​​the printed object. In practice, the area of ​​the printed object (foreground) can be marked as 1 and the background area as 0, thereby generating a background segmentation mask. This background segmentation mask can be a mask image used for background separation.

[0039] Step 5: Based on the aforementioned background segmentation mask, perform background removal processing on the aforementioned grayscale surface visual image to obtain a first background-removed surface visual image. In practice, the executing entity can perform a bitwise AND operation on the background segmentation mask and the grayscale surface visual image to obtain the first background-removed surface visual image. As an example, for each pixel in the grayscale surface visual image, the executing entity can determine the mask pixel corresponding to the background segmentation mask. Then, the pixel value corresponding to the mask pixel is converted to a grayscale pixel value and ANDed with the pixel value corresponding to the aforementioned pixel to obtain an updated pixel value (for example, as an example, the aforementioned converted grayscale pixel value mask pixel value can be 0, and the aforementioned pixel value to be updated can be 50, then the updated pixel value can be 0). Afterwards, the executing entity can update the updated pixel value to the pixel value of the pixel to update the grayscale surface visual image. Finally, the executing entity can determine the updated grayscale surface visual image as the first background-removed surface visual image.

[0040] The sixth step involves performing co-occurrence matrix (COM) feature extraction on the first background-removed surface visual image to obtain initial surface texture feature information. In practice, the execution entity can employ image texture analysis and extraction techniques based on gray-level co-occurrence matrix (GLCM) to perform COM feature extraction on the first background-removed surface visual image to obtain initial surface texture feature information. This initial surface texture feature information can be a feature vector representing texture features including contrast, homogeneity, energy, entropy, and correlation.

[0041] The sixth step involves performing local texture feature extraction processing on the first background-removed surface visual image to obtain local texture feature information. In practice, the execution entity can use Local Binary Pattern (LBP) technology to perform local texture feature extraction processing on the first background-removed surface visual image to obtain local texture feature information. This local texture feature information can be a feature vector representing the extracted local texture features that are insensitive to illumination.

[0042] Step 7: Based on the initial surface texture feature information and the local texture feature information mentioned above, generate surface texture feature information. In practice, feature fusion techniques (e.g., stitching techniques) can be used to fuse the initial surface texture feature information and the local texture feature information to obtain fused surface texture feature information.

[0043] Step 8: Based on the background segmentation mask described above, perform background removal processing on the denoised surface visual image to obtain a second background-removed surface visual image.

[0044] The ninth step involves extracting gloss characteristics from the second background-removed surface visual image to obtain gloss characteristic information. In practice, this can be achieved by calculating the area ratio and pixel intensity distribution of the highlight region in the second background-removed surface visual image, and then estimating the surface gloss characteristics using a reflection model based on these factors. The aforementioned gloss characteristic information can refer to information describing the surface gloss.

[0045] Step 10: Based on the aforementioned surface texture features and gloss characteristics, generate multi-dimensional fused feature information. In practice, the executing entity can employ feature fusion techniques (e.g., stitching techniques) to fuse the surface texture features and gloss characteristics, obtaining the fused information as multi-dimensional fused feature information.

[0046] Step 11: Input the aforementioned multi-dimensional fused feature information into the material feature recognition model to generate recognized surface feature information. This material feature recognition model can be a Convolutional Neural Network (CNN) model that takes the multi-dimensional fused feature information as input and outputs material feature recognition information (e.g., smoothness and transparency). Then, the executing entity can concatenate the material feature recognition information and the initial surface texture feature information, and determine the concatenated information as the recognized surface feature information (e.g., smoothness and transparency: {energy: 0.98, contrast: 0.05, homogeneity: 0.99}).

[0047] Step 12: Determine the above-mentioned surface feature information as image recognition information.

[0048] The above technical solution, combined with steps 102-103 and related content, serves as an inventive point of this disclosure, solving the technical problem of "poor print quality." Factors leading to poor print quality often include: material identification through images of the surface of the printed item; background interference in the acquired images; and the printing box being a closed or semi-closed space where uneven lighting and shadows easily occur, resulting in noise in the acquired images. Furthermore, the subtle differences in surface texture, gloss, and reflection between different materials make it difficult for traditional single-feature extraction methods to accurately distinguish between different material types, leading to inaccurate matching of subsequent printing parameters and consequently poor print quality. Solving these factors can improve print quality. To achieve this, firstly, the surface visual image is subjected to low-light environment noise reduction processing to obtain a denoised surface visual image. This suppresses and eliminates image noise caused by the low-light environment inside the printing box. Then, the denoised surface visual image is converted to grayscale to obtain a grayscale surface visual image. Next, the grayscale surface visual image is input into a pre-trained instance segmentation model to obtain information about the area of ​​the printed item. Therefore, the area containing the printed object in the image can be accurately located and separated, obtaining the printed object area information. Then, based on the printed object area information, a background segmentation mask is generated. Next, based on the background segmentation mask, the grayscale surface visual image is subjected to background removal processing to obtain a first background-removed surface visual image. This removes cluttered backgrounds, resulting in the first background-removed surface visual image. Co-occurrence matrix feature extraction processing is performed on the first background-removed surface visual image to obtain initial surface texture feature information. This captures the spatial relationships and macroscopic texture patterns between pixel pairs in the image, quantifies macroscopic attributes such as material roughness and contrast, and obtains initial surface texture feature information. Next, local texture feature extraction processing is performed on the first background-removed surface visual image to obtain local texture feature information. This captures the local texture feature information of the image. Then, based on the initial surface texture feature information and the local texture feature information, surface texture feature information is generated. This fuses macroscopic and microscopic texture information to form a comprehensive and robust texture feature representation, i.e., surface texture feature information. Next, based on the aforementioned background segmentation mask, background removal processing is performed on the denoised surface visual image to obtain a second background-removed surface visual image. Gloss characteristic extraction processing is then performed on the second background-removed surface visual image to obtain gloss characteristic information. Next, based on the aforementioned surface texture feature information and gloss characteristic information, multi-dimensional fused feature information is generated. Thus, texture information characterizing the surface's physical structure and gloss information characterizing the surface's optical properties can be combined to form a comprehensive material feature description, resulting in multi-dimensional fused feature information.Then, the aforementioned multi-dimensional fused feature information is input into the material feature recognition model to generate surface feature information for recognition. This surface feature information is then defined as image recognition information. Because it extracts and fuses multi-dimensional key features such as texture and gloss from a clean image after removing background and noise, it effectively overcomes the problems of background interference, noise, and insufficient discrimination power of single features, generating more accurate image recognition information. Based on this more accurate image recognition information, more accurate item attribute information is constructed. Combining this with steps 102-103, more precise adaptive printing parameters can be matched to the printed item based on the more accurate item attribute information, improving print quality.

[0049] Step 102: Determine the item identifier of the item to be printed based on the item attribute information.

[0050] In some embodiments, the executing entity may determine the item identifier of the printed item based on the item attribute information. The item identifier may be the material name or item name of the printed item (e.g., acrylic sheet, phone case).

[0051] In some optional implementations of certain embodiments, the aforementioned executing entity may determine the item identifier of the item to be printed based on item attribute information through the following steps: The first step involves performing a fusion cross-validation recognition process on the image recognition information, near-field communication recognition information, and weight information included in the aforementioned item attribute information to obtain the item identifier. In practice, the executing entity can use information fusion algorithms (such as decision trees, weighted voting, or simple rule engines) to perform the fusion cross-validation recognition process on the image recognition information, near-field communication recognition information, and weight information included in the aforementioned item attribute information to obtain the item identifier. The aforementioned image recognition information includes surface feature information, and the aforementioned near-field communication recognition information includes the identifier to be verified. The aforementioned surface feature information can be information representing the surface features of the printed item (e.g., smoothness and transparency: energy: 0.98, contrast: 0.05, homogeneity: 0.99).

[0052] In some optional implementations of certain embodiments, the aforementioned executing entity may perform the following fusion cross-validation recognition process on the image recognition information, near-field communication recognition information, and weight information included in the aforementioned item attribute information to obtain the item identifier: The first step is to identify the identification identifiers to be verified included in the aforementioned near-field communication identification information.

[0053] The second step is to obtain at least one verification identification surface feature information corresponding to the aforementioned identification mark to be verified. Each of the above-mentioned verification identification surface feature information can be a pre-set identification surface feature information corresponding to the identification mark to be verified (for example, the verification identification surface feature information can be smooth and transparent: {energy: 0.98, contrast: 0.05, homogeneity: 0.99}).

[0054] Thirdly, in response to determining that the similarity between the image recognition information (including the surface feature information) and one of the at least one verified surface feature information is greater than a preset similarity, a first preset positive score is determined as the first score corresponding to the identification identifier to be verified. The similarity can be cosine similarity. The first preset positive score can be a preset value greater than zero, for example, 7 points.

[0055] Fourth step: In response to determining that there is no verification identification surface feature information among the at least one verification identification surface feature information that has a similarity greater than a preset similarity with the identification surface feature information, a first preset negative score is determined as the first score corresponding to the identification identifier to be verified. The first preset negative score can be a preset value less than zero, for example, -3 points.

[0056] The fifth step is to obtain the weight range information corresponding to the identification mark to be verified mentioned above. For example, the weight range information corresponding to "acrylic" could be "130g-150g".

[0057] Step 6: In response to determining that the weight represented by the aforementioned weight information is within the range represented by the aforementioned weight range information, a second preset positive score is determined as the second score corresponding to the aforementioned identification identifier to be verified. The aforementioned second preset positive score can be a preset value greater than zero, for example, 8 points.

[0058] Step 7: In response to determining that the weight represented by the aforementioned weight information is not within the range represented by the aforementioned weight range information, a second preset negative score is determined as the negative score corresponding to the aforementioned identification identifier to be verified. The aforementioned second preset negative score can be a preset value less than zero.

[0059] Step 8: Based on the determined first and second scores, generate a comprehensive score. In practice, the implementing entity can determine the comprehensive score as the sum of the first and second scores.

[0060] Step 9: In response to determining that the comprehensive score is greater than or equal to the preset score, the above-mentioned identification mark to be verified is identified as the item mark.

[0061] Step 103: Match the adaptive printing parameter information corresponding to the item identifier.

[0062] In some embodiments, the execution entity may match adaptive printing parameter information corresponding to the item identifier. The adaptive printing parameter information may be printing parameter information corresponding to a preset printing scheme corresponding to the item identifier. The printing parameter information includes various printing parameters. These printing parameters include, but are not limited to, at least one of the following: droplet size, jetting frequency, UV curing lamp power and irradiation time, and distance between the print head and the item surface.

[0063] In some optional implementations of certain embodiments, the aforementioned execution entity may match the adaptive printing parameter information corresponding to the aforementioned item identifier through the following steps: The first step is to query the printing parameter information corresponding to the aforementioned item identifier from the preset printing parameter information database. This database stores the applicable printing parameters for different items corresponding to different item identifiers. Each printing parameter is represented by a corresponding item identifier, indicating the preset printing scheme applicable to that item. The preset printing parameter information database can be a relational database storing these printing parameters.

[0064] The second step is to determine the retrieved printing parameter information as adaptive printing parameter information.

[0065] Step 104: Control the printing device to print the item based on adaptive printing parameter information.

[0066] In some embodiments, the aforementioned execution entity can control the printing device to print the aforementioned item based on the aforementioned adaptive printing parameter information. The printing device can be a UV printer. In practice, the aforementioned execution entity can send the adaptive printing parameter information and a preset image (i.e., the print image) to the UV printer execution module (physical hardware: including traditional components such as the print head, UV curing lamp, and motion platform, but whose operating parameters are controlled by the central processing unit module in the UV printer), and drive it to work together to print the preset image onto the surface of the item. Specifically, as an example, the aforementioned execution entity can perform a composite processing of the adaptive printing parameter information and the preset image input by the user. This processing includes: rasterization: converting the user-input vector or pixel image into a two-dimensional bitmap of ink droplets (e.g., color model: C, M, Y, K, W) that the printer nozzle needs to eject at each physical location. Job composition: if the adaptive parameters specify that a white primer needs to be printed first, a white primer print layer based on the user image outline is automatically generated, and the complete print job is planned as a logical sequence of "printing the white base layer first, then printing the color image layer". Parameter Integration: The various parameters included in the adaptive printing parameter information (such as droplet size and jet frequency) are integrated and encoded into the final control command stream sent to the printer. This control command stream is then sent to the various sub-controllers of the UV printing execution module, driving them to work collaboratively: Motion System Control: Commands are sent to the motion controller to drive the motor, controlling the print head or platform to move precisely at the specified printing speed and print head height. Inkjet System Control: The rasterized image bitmap data is synchronously sent to the print head driver. Based on the droplet size and jet frequency parameters integrated into the command stream, the driver controls specific nozzles to eject droplets of the specified color of ink at precise moments. Curing System Control: Commands are sent to the UV lamp controller to immediately activate the UV lamp according to the specified UV lamp power and irradiation time after the corresponding printed layer is completed, curing the ink.

[0067] The above-described embodiments of this disclosure have the following beneficial effects: the adaptive printing method based on multimodal recognition in some embodiments of this disclosure improves printing quality and printing production efficiency. Specifically, the reasons for poor printing quality and low printing production efficiency are: by identifying items from a single information source, the material, surface characteristics, and identity information of the object cannot be comprehensively judged, resulting in poor accuracy of the identification results, which in turn leads to poor adaptability of the matching printing parameters and poor printing quality (e.g., ink bleeding, insufficient adhesion, color deviation, and other printing quality problems). At the same time, the printing production efficiency is low due to reliance on manual experience to adjust parameters. Based on this, the adaptive printing method based on multimodal recognition in some embodiments of this disclosure first performs multimodal recognition processing on the item to be printed to obtain item attribute information. Thus, multidimensional attribute information of the item to be printed can be obtained. Next, based on the item attribute information, the item identifier of the item to be printed is determined. Thus, comprehensive analysis and judgment can be performed based on multidimensional attribute information, avoiding the identification bias caused by a single information source and improving the accuracy and reliability of item identifier (item type or identity identifier). Next, the adaptive printing parameter information corresponding to the aforementioned item identifier is matched. Thus, based on the accurately determined item identifier, the optimal set of printing parameters for that type of item is matched, i.e., adaptive printing parameter information adapted to the specific material and surface characteristics of the item being printed. Then, the printing equipment is controlled to print the item based on the aforementioned adaptive printing parameter information. Therefore, printing is performed according to the matched printing parameter information adapted to the specific material and surface characteristics of the item being printed, reducing printing defects caused by parameter mismatches and improving print quality. Furthermore, because the entire process from identifying and matching adaptive printing parameter information to executing printing is completed without manual intervention, production downtime and reliance on operator experience are significantly reduced, thus improving printing production efficiency. Figure 4 These are internal product diagrams of some embodiments of the printing apparatus according to this disclosure. Figure 5 It is an internal test diagram of the printing device according to this disclosure during the printing process; Figure 6 This is an internal test printout of an adaptive printing method based on multimodal recognition, which is used in accordance with this disclosure and printed by a printing device.

[0068] In addressing the technical problems mentioned above by adopting technical solutions, the application scenario—an adaptive printing system based on closed-loop quality feedback—often presents the following technical challenges: After printing using adaptive printing parameters, there is typically a lack of print quality inspection of the printed items. Furthermore, when print quality defects occur, it is difficult to correlate specific color deviations, insufficient clarity, or physical defects with the adopted adaptive printing parameters, failing to form a closed-loop data chain of "parameter settings - print results." This results in the print parameter database remaining stagnant, causing printing to remain at its initial level and hindering continuous improvement in print quality. Therefore, this application scenario requires the following characteristics: suitability for quality inspection of print effects, and the ability to update the print parameter information database based on the quality inspection results.

[0069] In some optional implementations of certain embodiments, the aforementioned execution entity may further perform the following steps: After printing is complete, control the printing platform carrying the printed item to move to the preset imaging area.

[0070] The first step is to acquire an image of a preset imaging area using an image acquisition device as the printed result. This image acquisition device can be a camera mounted above the preset imaging area.

[0071] The second step is to obtain the print image input by the user.

[0072] The third step involves performing color fidelity detection processing on the printed image and the printed effect image to obtain color fidelity detection information. In practice, the executing entity can convert the printed image and the printed effect image to the Lab color space. Then, after converting both images to grayscale, it performs pixel-by-pixel color difference calculation (e.g., calculating the difference between the pixel value of each pixel in the grayscale-processed printed image and the corresponding pixel value in the grayscale-processed printed effect image (e.g., subtracting the pixel value of the corresponding pixel in the grayscale-processed printed effect image from the pixel value in the grayscale-processed printed image)). This yields various color difference values. Next, the executing entity can determine the average of these color difference values ​​as the color difference mean. Finally, the executing entity can use this color difference mean as the color fidelity detection information.

[0073] The fourth step involves performing a sharpness detection process on the printed image to obtain sharpness detection information. In practice, the execution entity can use image sharpness detection techniques (such as gradient functions) to perform sharpness detection on the printed image and obtain sharpness detection information. Specifically, firstly, the gradient of each pixel in the printed image can be calculated (for example, using the Sobel operator). Then, the mean value of each gradient obtained for each pixel is determined as the sharpness as the sharpness detection information.

[0074] The fifth step involves processing the printed image to detect printing defects, obtaining printing defect detection information. In practice, pre-trained defect detection models (such as YOLO or U-Net) can be used to identify whether the printed image has obvious printing defects (such as ink splatter, stringing, poor curing, etc.), thus obtaining printing defect detection information. This printing defect detection information can be textual information indicating whether the printed image has obvious printing defects (e.g., obvious printing defects exist: True or no obvious printing defects: False).

[0075] Step 6: Based on the aforementioned color fidelity detection information, sharpness detection information, and printing defect detection information, a print quality score and deviation report are generated. In practice, in response to determining that the printing defect detection information indicates obvious printing defects in the printed image, the executing entity can determine a first preset defect detection score (e.g., -5) as the defect detection score. In response to determining that the printing defect detection information indicates no obvious printing defects in the printed image, the executing entity can determine a first preset defect detection score (e.g., +5) as the defect detection score. Next, the executing entity can determine the negative of the mean color difference represented by the color fidelity detection information as the color fidelity value. Afterward, the executing entity can determine the print quality score based on the weighted average of the defect detection score, the color fidelity value, and the sharpness represented by the sharpness detection information, using preset weights for the defect detection score, the color fidelity value, and the sharpness score. In response to determining that the mean color difference represented by the color fidelity detection information is greater than a preset value, text information indicating insufficient color saturation in the printed image is determined as color deviation information. In response to the determination that the sharpness represented by the sharpness detection information is less than a preset sharpness, the text information representing the poor sharpness of the printed image is identified as sharpness deviation information. In response to the determination that the printing defect detection information represents obvious printing defects in the printed image, the text information representing obvious printing defects in the printed image is identified as printing defect deviation information. One or more of the color deviation information, sharpness deviation information, and printing defect deviation information are identified as a deviation report.

[0076] Step 7: In response to determining that the above print quality score is less than a preset threshold, the above adaptive printing parameter information, the above print quality score, and the above deviation report are identified as the samples to be optimized corresponding to the item identifier.

[0077] Step 8: Query the number of historical samples to be optimized corresponding to the above-mentioned item identifier from the preset printing parameter information database, and in response to determining that the above-mentioned number is greater than the preset number threshold, obtain each historical sample to be optimized corresponding to the above-mentioned item identifier from the preset printing parameter information database.

[0078] Step 9: Based on the aforementioned historical samples to be optimized and the samples to be optimized, generate optimized printing parameter information corresponding to the aforementioned item identifier, and update the optimized printing parameter information to the preset printing parameter information library to update the preset printing parameter information library. For example, if multiple historical samples to be optimized report deviations all indicate "insufficient color saturation," and these samples all use similar "ink density" parameters, then generate a new "ink density" parameter suggestion value for that item identifier (e.g., increasing the original value by 5%). This new "ink density" parameter suggestion value (e.g., increasing the original value by 5%) is sent to the operator. After the operator confirms, the generated parameter suggestion value is added to the adaptive printing parameter information, overwriting the corresponding parameter in the adaptive printing parameter information to update the adaptive printing parameter information. Then, the updated adaptive printing parameter information corresponding to the item identifier is updated as optimized printing parameter information to the preset printing parameter information library, overwriting the original printing parameter information corresponding to the item identifier in the preset printing parameter information library. Optionally, for unrealized future plans: the system can introduce a reinforcement learning model, treat the adjustment of printing parameters as the action of the agent, and use the printing quality score as a reward signal, so that the model can learn the optimal parameter strategy autonomously through a large number of virtual or real printing experiments, and achieve fully adaptive parameter tuning.

[0079] Step 10: Perform adaptive printing based on the updated preset print parameter information library.

[0080] The above-described technical solution and its related content, as an inventive point of this disclosure, solve the technical problem of "printing remaining at the initial level for a long time, unable to continuously improve print quality." Factors causing printing to remain at the initial level for a long time, unable to continuously improve print quality, are often as follows: Adaptive printing systems based on closed-loop quality feedback are often accompanied by the following technical problems: After printing is completed using adaptive printing parameters, there is usually a lack of print quality detection on the printed items, and when print quality defects occur, it is difficult to correlate specific color deviations, insufficient clarity, or physical defects with the adopted adaptive printing parameters, failing to form a closed-loop data chain of "parameter settings - printing results," causing the print parameter library to stagnate for a long time, resulting in printing remaining at the initial level and unable to continuously improve print quality. If the above factors are solved, the effect of continuously improving print quality can be achieved. To achieve this effect, firstly, after printing is completed, the printing platform carrying the printed items is moved to a preset imaging area. An image of the preset imaging area is acquired through an image acquisition device as a print effect image. Thus, an image reflecting the true effect of this printing can be obtained. Next, the print image input by the user is acquired. Then, based on the printed image, color fidelity testing is performed on the printed effect image to obtain color fidelity testing information. This quantifies the degree of color reproduction between the printed product and the original image, yielding color fidelity testing information. Next, sharpness testing is performed on the printed effect image to obtain sharpness testing information. This provides data on sharpness for evaluating print quality, yielding sharpness testing information. Then, printing defect detection is performed on the printed effect image to obtain printing defect detection information. This detects physical printing defects such as ink splatter, streaking, and striations, yielding printing defect detection information. Finally, based on the color fidelity testing information, sharpness testing information, and printing defect detection information, a print quality score and deviation report are generated. This integrates multiple quality indicators into a single print quality score and generates a deviation report specifying particular problems (such as "insufficient color saturation" or "defects present"). Subsequently, in response to determining that the print quality score is less than a preset threshold, the adaptive print parameter information, the print quality score, and the deviation report are identified as samples to be optimized corresponding to the item identifier. Next, the number of historical samples to be optimized corresponding to the item identifier is retrieved from the preset print parameter information database, and in response to determining that the number is greater than a preset number threshold, each historical sample to be optimized corresponding to the item identifier is obtained from the preset print parameter information database. Finally, based on each historical sample to be optimized and the sample to be optimized, optimized print parameter information corresponding to the item identifier is generated, and the optimized print parameter information is updated to the preset print parameter information database to update the preset print parameter information database.This allows us to determine if there is sufficient historical data for analysis. If so, historical samples for optimization are retrieved. Based on the information from both historical and current samples, optimized printing parameters corresponding to the item identifier are generated and updated in the printing parameter database. This ensures the database is continuously updated, allowing for ongoing improvement of printing parameters and forming a dynamic optimization mechanism to enhance future print quality. Finally, adaptive printing is performed based on the updated preset printing parameter database. Thus, by using the updated database and performing adaptive printing, each print job is based on the latest optimized parameters, achieving continuous improvement in print quality.

[0081] Further reference Figure 2 As an implementation of the methods shown in the figures, this disclosure provides some embodiments of an adaptive printing apparatus based on multimodal recognition, which are similar to... Figure 1 Corresponding to the method embodiments shown, the device can be specifically applied to various electronic devices.

[0082] like Figure 2 As shown, an adaptive printing apparatus 200 based on multimodal recognition in some embodiments includes: a multimodal recognition processing unit 201, a determining unit 202, a matching unit 203, and a printing unit 204. The multimodal recognition processing unit is configured to perform multimodal recognition processing on the item to be printed to obtain item attribute information; the determining unit is configured to determine an item identifier of the item to be printed based on the item attribute information; the matching unit is configured to match adaptive printing parameter information corresponding to the item identifier; and the printing unit is configured to control a printing device to print the item to be printed based on the adaptive printing parameter information.

[0083] It is understandable that the units described in the device 200 are related to the reference. Figure 1 The steps in the method described above correspond to each other. Therefore, the operations, features, and beneficial effects described above for the method also apply to the device 200 and the units contained therein, and will not be repeated here.

[0084] The following is for reference. Figure 3 It shows a schematic diagram of the structure of an electronic device 300 suitable for implementing some embodiments of the present disclosure. Figure 3 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments of this disclosure.

[0085] like Figure 3As shown, electronic device 300 may include a printing device and a processing unit (e.g., a central processing unit, a graphics processor, etc.) 301, which can perform various appropriate actions and processes according to a program stored in read-only memory (ROM) 302 or a program loaded from storage device 308 into random access memory (RAM) 303. The RAM 303 also stores various programs and data required for the operation of electronic device 300. The processing unit 301, ROM 302, and RAM 303 are interconnected via bus 304. Input / output (I / O) interface 305 is also connected to bus 304.

[0086] Typically, the following devices can be connected to I / O interface 305: input devices 306 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 307 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 308 including, for example, magnetic tapes, hard disks, etc.; and communication devices 309. Communication device 309 allows electronic device 300 to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 3 An electronic device 300 with various devices is shown; however, it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed alternatively. Figure 3 Each box shown can represent a device or multiple devices as needed.

[0087] In particular, according to some embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, some embodiments of this disclosure include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication device 309, or installed from storage device 308, or installed from ROM 302. When the computer program is executed by processing device 301, it performs the functions defined in the methods of some embodiments of this disclosure.

[0088] It should be noted that, in some embodiments of this disclosure, the computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium may be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In some embodiments of this disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In some embodiments of this disclosure, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0089] In some implementations, clients and servers can communicate using any currently known or future-developed network protocol such as HTTP (Hypertext Transfer Protocol) and can interconnect with digital data communication (e.g., communication networks) of any form or medium. Examples of communication networks include local area networks (“LANs”), wide area networks (“WANs”), the Internet (e.g., the Internet of Things), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future-developed networks.

[0090] A computer-readable medium may be included in an electronic device or may exist independently without being assembled into the electronic device. The computer-readable medium carries one or more programs that, when executed by the electronic device, cause the electronic device to: perform multimodal recognition processing on the item to be printed to obtain item attribute information; determine an item identifier for the item to be printed based on the item attribute information; match adaptive printing parameter information corresponding to the item identifier; and control a printing device to print the item to be printed based on the adaptive printing parameter information.

[0091] Computer program code for performing operations of some embodiments of this disclosure can be written in one or more programming languages ​​or a combination thereof. Programming languages ​​include object-oriented programming languages—such as Java, Smalltalk, and C++—and conventional procedural programming languages—such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0092] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0093] The units described in some embodiments of this disclosure can be implemented in software or hardware. The described units can also be housed in a processor; for example, a processor may be described as including a multimodal recognition processing unit, a determining unit, a matching unit, and a printing unit. The names of these units do not necessarily limit the specific unit; for example, a multimodal recognition processing unit may also be described as "a unit that performs multimodal recognition processing on a printed item to obtain item attribute information."

[0094] The functions described above in this document can be performed at least in part by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), system-on-a-chip (SoCs), complex programmable logic devices (CPLDs), and so on.

[0095] The above description is merely a selection of preferred embodiments of this disclosure and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in the embodiments of this disclosure is not limited to technical solutions formed by specific combinations of technical features, but should also cover other technical solutions formed by arbitrary combinations of technical features or their equivalents without departing from the inventive concept. For example, technical solutions formed by substituting features with (but not limited to) technical features with similar functions disclosed in the embodiments of this disclosure.

Claims

1. An adaptive printing method based on multimodal recognition, comprising: Multimodal recognition processing is performed on the printed item to obtain the item attribute information; Based on the item attribute information, determine the item identifier of the item to be printed; Match the adaptive printing parameter information corresponding to the item identifier; The printing device is controlled to print the item based on the adaptive printing parameter information.

2. The method according to claim 1, wherein, The process of performing multimodal recognition on the printed item to obtain item attribute information includes: At least two items of object perception and recognition information are collected from the printed item to form item attribute information, wherein the at least two items of object perception and recognition information include at least two of the following: image recognition information, near-field communication recognition information, and weight information.

3. The method according to claim 2, wherein, The collection of at least two items of object perception and recognition information of the printed item includes: Acquire a visual image of the surface of the printed item, and generate image recognition information based on the visual image of the surface; Read the NFC tag embedded on the printed item or its carrier tray to obtain near-field communication identification information; Collect the weight information of the printed item.

4. The method according to claim 3, wherein, The step of generating image recognition information based on the surface visual image includes: The surface visual image is subjected to feature extraction processing to obtain the visual feature information of the printed item; Image recognition information is generated based on the visual feature information of the printed item.

5. The method according to claim 1, wherein, The item attribute information includes image recognition information, near-field communication recognition information, and weight information, as well as the item identifier determined based on the item attribute information, including: The image recognition information, near-field communication recognition information, and weight information included in the item attribute information are subjected to the following fusion cross-validation recognition process to obtain the item identifier.

6. The method according to claim 1, wherein, The adaptive printing parameter information matching the item identifier includes: Retrieve the printing parameter information corresponding to the item identifier from the preset printing parameter information database; The retrieved printing parameter information is determined as adaptive printing parameter information.

7. The method according to claim 5, wherein, The image recognition information includes surface feature information, and the near-field communication recognition information includes an identification identifier to be verified. The following fusion and cross-validation recognition process is performed on the image recognition information, near-field communication recognition information, and weight information included in the item attribute information to obtain the item identifier: Determine the identification identifier to be verified included in the near-field communication identification information; Obtain at least one verification identification surface feature information corresponding to the identification identifier to be verified; In response to determining that the similarity between the image recognition information, including the identification surface feature information, and one of the at least one verification identification surface feature information is greater than a preset similarity, a first preset positive score is determined as the first score corresponding to the identification identifier to be verified; In response to determining that there is no verification identification surface feature information in the at least one verification identification surface feature information that has a similarity greater than a preset similarity with the identification surface feature information, the first preset negative score is determined as the first score corresponding to the identification identifier to be verified; Obtain the weight range information corresponding to the identification identifier to be verified; In response to determining that the weight represented by the weight information is within the range represented by the weight range information, the second preset positive score is determined as the second score corresponding to the identification identifier to be verified; In response to determining that the weight represented by the weight information is not within the range represented by the weight range information, the second preset negative score is determined as the negative score corresponding to the identification identifier to be verified; A comprehensive score is generated based on the determined first and second scores; In response to determining that the comprehensive score is greater than or equal to the preset score, the identification mark to be verified is identified as the item mark.

8. An adaptive printing device based on multimodal recognition, comprising: The multimodal recognition processing unit is configured to perform multimodal recognition processing on the printed item to obtain item attribute information; The determination unit is configured to determine the item identifier of the item to be printed based on the item attribute information; The matching unit is configured to match adaptive printing parameter information corresponding to the item identifier; The printing unit is configured to control the printing device to print the item based on the adaptive printing parameter information.

9. An electronic device, comprising: Printing equipment, one or more processors; A storage device on which one or more programs are stored; When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1 to 7.

10. A computer-readable medium having a computer program stored thereon, wherein, When the program is executed by the processor, it implements the method as described in any one of claims 1 to 7.

Images