Image vectorization garment entry method and system

Abstract

The invention relates to the technical field of garment manufacturing, in particular to an image vectorization garment input method and system, and the method comprises the steps: scanning a garment to obtain a garment digital image; identifying a garment contour in the garment digital image by using an image processing technology; converting the garment contour into vector graphic data; and generating a digital model of the clothing according to the vector graphic data. Through the steps, a costume designer and a costume manufacturer efficiently and accurately convert physical costumes into editable and extensible digital assets, so that the costume design and production process is greatly simplified, and the efficiency and the quality are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of clothing manufacturing, and in particular to an image vectorized clothing entry method and system. Background Art

[0002] Clothing entry refers to the process of recording relevant information about clothing into a computer system in a digital form. This process usually includes a detailed description of the clothing, such as style, color, size, material, brand, price, etc., and contains high-quality multi-angle pictures or video materials. Clothing entry is widely used in e-commerce platforms, retail management systems and inventory control systems, which helps to improve product management efficiency, optimize customer shopping experience and support precision marketing strategies.

[0003] The existing method requires manual scanning of clothing and parameter modeling, which has low work efficiency and high cost. Summary of the invention

[0004] The purpose of the present invention is to provide an image vectorized clothing entry method and system, aiming to efficiently and accurately convert physical clothing into editable and scalable digital assets, thereby greatly simplifying the process of clothing design and production.

[0005] To achieve the above-mentioned object, in a first aspect, the present invention provides an image vectorized clothing entry method, comprising scanning clothing to obtain a digital image of the clothing;

[0006] Using image processing technology to identify the garment contour in the garment digital image;

[0007] Converting the garment outline into vector graphics data;

[0008] A digital model of the garment is generated according to the vector graphic data.

[0009] The specific steps of scanning the clothing to obtain the digital image of the clothing include:

[0010] Iron the garment flat before scanning to ensure there are no wrinkles;

[0011] Adjust the scanner to acquire garment images according to the resolution and color mode of the desired images;

[0012] The clothing images are stitched into a complete clothing digital image.

[0013] The specific steps of stitching the clothing images into a complete clothing digital image include:

[0014] Adjust the curve smoothness of the clothing image and remove the overlapping parts;

[0015] The clothing image is segmented and adjusted according to the required size and shape to obtain clothing units;

[0016] All clothing units are spliced ​​together to obtain a clothing digital image.

[0017] The specific steps of using image processing technology to identify the clothing outline in the clothing digital image include:

[0018] Converting digital images of clothing into grayscale images;

[0019] Apply filters to remove noise from grayscale images;

[0020] Use the Sobel operator to calculate the gradient of the grayscale image and find the edge of the image;

[0021] The contour that best matches the garment contour is selected according to the characteristic factors of the contour, wherein the characteristic factors include area, perimeter, center point, minimum circumscribed rectangle and minimum circumscribed circle.

[0022] The specific steps of using the Sobel operator to calculate the gradient of the grayscale image and find the edge of the image include:

[0023] Normalize the pixel value range to -1 to 1;

[0024] Use the two matrices of the Sobel operator to perform convolution operations with each pixel in the image to obtain the gradient components in the horizontal and vertical directions;

[0025] Calculate the gradient magnitude of each pixel;

[0026] Calculate the gradient direction of each pixel;

[0027] Divide pixels into multiple direction groups according to the gradient direction;

[0028] For each pixel, check whether its gradient magnitude is greater than the gradient magnitudes of the pixels along the gradient direction in its two neighborhoods;

[0029] Set a high threshold and a low threshold. Pixels above the high threshold are considered strong edge points; pixels below the low threshold are considered background.

[0030] The specific steps of selecting the contour that best matches the clothing contour according to the characteristic factors of the contour include:

[0031] For each contour, calculate the contour characteristics

[0032] Set a matching threshold range to exclude contours outside the range.

[0033] The specific steps of converting the clothing outline into vector graphics data include:

[0034] Use OpenCV to extract contours and convert them into vector paths;

[0035] Save vector paths as SVG files.

[0036] The specific steps of using OpenCV to extract the contour and convert it into a vector path include:

[0037] Represent each contour point with coordinates;

[0038] Calculate the distance between adjacent points and sort them based on the distance;

[0039] Use the Douglas-Peucker algorithm to reduce the number of contour points;

[0040] The contour points are fitted using cubic Bezier curves to create smooth curve segments to obtain vector paths.

[0041] In a second aspect, the present invention provides an image vectorized clothing entry system, comprising a clothing scanning module, a contour recognition module, a contour conversion module and a model generation module;

[0042] The clothing scanning module is used to scan the clothing to obtain a digital image of the clothing;

[0043] The outline recognition module is used to recognize the clothing outline in the clothing digital image using image processing technology;

[0044] The outline conversion module is used to convert the clothing outline into vector graphics data;

[0045] The model generation module is used to generate a digital model of clothing according to the vector graphics data.

[0046] The invention discloses an image vectorization clothing entry method and system, comprising: scanning clothing to obtain a digital image of clothing; using image processing technology to identify the clothing outline in the digital image of clothing; converting the clothing outline into vector graphics data; and generating a digital model of clothing according to the vector graphics data. The physical clothing to be entered is placed on a dedicated scanning platform. The clothing is fully scanned using a high-resolution scanning device to capture all the details and features of the clothing. The scanning process includes multi-angle shooting to ensure that a complete image of each part of the clothing can be obtained. After the scanning is completed, multiple pictures are synthesized into one or more high-quality digital images of clothing by using image stitching and other technologies. Computer vision technology, such as edge detection algorithm (such as Canny edge detection), threshold segmentation or other advanced image segmentation technology is used to identify the key contour lines in the digital image of clothing. Image enhancement technology, such as contrast enhancement, noise reduction processing, etc., is also required to be used to more accurately identify the details of the clothing. At this stage, machine learning or deep learning models, such as convolutional neural network (CNN), are also used to improve the accuracy and robustness of contour recognition. The clothing outline identified in the previous step is converted into vector graphics data by image vectorization technology. This process usually involves converting pixel-level contour lines into curves expressed in mathematical form, such as Bezier curves or spline curves. The vectorized data not only retains the information of the original contour, but also facilitates subsequent editing and enlargement without distortion. Based on the vectorized clothing contour data, a three-dimensional (3D) or two-dimensional (2D) clothing digital model is constructed. This process requires the use of professional CAD software or specific clothing design software. The digital model contains information such as the material properties, color, and texture of the clothing, making the final model as close to the effect of the physical clothing as possible. The generated digital model can be used in a variety of application scenarios such as virtual try-on, online display, design modification, etc.

[0047] Through the above steps, clothing designers and manufacturers can efficiently and accurately transform physical garments into editable and scalable digital assets, which greatly simplifies the clothing design and production process and improves efficiency and quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0049] Figure 1 It is a flow chart of a method for entering clothing into image vectorization according to the first embodiment of the present invention.

[0050] Figure 2 It is a flow chart of scanning clothing to obtain a digital image of clothing according to the first embodiment of the present invention.

[0051] Figure 3 This is a flow chart of splicing clothing images into a complete clothing digital image according to the first embodiment of the present invention.

[0052] Figure 4 The flowchart of the first embodiment of the present invention is to recognize the clothing outline in the clothing digital image by using the image processing technology.

[0053] Figure 5 This is a flow chart of using the Sobel operator to calculate the gradient of a grayscale image and find the edge of the image according to the first embodiment of the present invention.

[0054] Figure 6 It is a flow chart of selecting a contour that best matches a clothing contour according to characteristic factors of the contour according to the first embodiment of the present invention.

[0055] Figure 7 It is a flowchart of converting the clothing outline into vector graphic data according to the first embodiment of the present invention.

[0056] Figure 8 It is a flowchart of using OpenCV to extract contours and convert them into vector paths according to the first embodiment of the present invention. DETAILED DESCRIPTION

[0057] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention.

[0058] First embodiment

[0059] See also Figures 1 to 8 The present invention provides an image vectorized clothing input method, comprising:

[0060] S101 scans the clothing to obtain a digital image of the clothing;

[0061] The specific steps include:

[0062] S201 irons the garments flat before scanning to ensure there are no wrinkles;

[0063] Choose a flat and stable work surface and use a professional iron and flattening tool to remove any creases or wrinkles from the garment. This step is crucial because any wrinkles will show up in the final digital image and affect the subsequent design and production process.

[0064] Choose the appropriate ironing temperature and method according to the material of the clothing to avoid damaging the fabric.

[0065] S202 adjusts the scanner to obtain the clothing image according to the resolution and color mode of the required image;

[0066] Determine the appropriate resolution (such as 300 DPI for high-quality printing) based on the final use (such as for printing, web presentation, etc.). At the same time, choose the correct color mode (such as RGB for screen display, CMYK for printing). Make sure the scanner is properly calibrated to obtain accurate color reproduction.

[0067] S203 stitches the clothing images into a complete clothing digital image.

[0068] The specific steps include:

[0069] S301 adjusts the curve smoothness of the clothing image and deletes the overlapping parts;

[0070] Use professional image processing software (such as Adobe Photoshop or GIMP) to adjust the curve smoothness manually or through the automatic function to ensure that the image edges are clear and without jagged edges. If there are multiple scanned images, you need to carefully compare and crop the overlapping parts to ensure a smooth transition between the parts.

[0071] S302: segmenting the clothing image and adjusting it according to the required size and shape to obtain clothing units;

[0072] Divide the image into different parts or components according to the design requirements, such as sleeves, necklines, pockets, etc. Use the Warp tool to adjust the size and shape of each part to make it conform to the actual size proportion.

[0073] S303: All clothing units are spliced ​​together to obtain a clothing digital image.

[0074] All adjusted clothing units are stitched together according to their original positions to form a complete digital image of the clothing. After stitching is completed, the overall brightness, contrast, etc. are further adjusted to ensure the final image quality.

[0075] Through the above steps, high-quality digital images of clothing can be obtained to provide support for subsequent design, production and marketing.

[0076] S102 uses image processing technology to identify clothing contours in the clothing digital image;

[0077] The specific steps include:

[0078] S401 converts the clothing digital image into a grayscale image;

[0079] Use image processing software or programming languages ​​(such as Python's OpenCV library) to convert color images into grayscale images. This is because grayscale images simplify the processing flow, reduce the amount of calculation, and facilitate subsequent edge detection.

[0080] Usually the weighted average method is used, for example, Y = 0.299R + 0.587G +

[0081] 0.114BY=0.299R+0.587G+0.114B, where R, G, BR, G, and B represent the intensity values ​​of the red, green, and blue color channels respectively.

[0082] S402 applies a filter to remove noise in the grayscale image;

[0083] Select the appropriate filter type, such as mean filter, Gaussian filter or median filter, to reduce the random noise in the image. Through convolution operation, the filter template is applied to each pixel in the image to achieve the effect of smoothing the image.

[0084] S403 uses the Sobel operator to calculate the gradient of the grayscale image and find the edge of the image;

[0085] The specific steps include:

[0086] S501 normalizes the pixel value range to -1 to 1;

[0087] Normalization: Normalize the pixel values ​​of the grayscale image from the range of 0 to 255 to the range of -1 to 1 to meet the calculation requirements of the Sobel operator.

[0088] S502 uses two matrices of the Sobel operator to perform convolution operations on each pixel in the image to obtain gradient components in the horizontal direction and the vertical direction;

[0089] The two matrices of the Sobel operator are used to perform convolution operations with each pixel in the image to obtain the gradient components in the horizontal and vertical directions.

[0090] The horizontal gradient component is convolved using the following matrix:

[0091]

[0092] The vertical gradient component is convolved using the following matrix:

[0093]

[0094] Convolve the above matrix with each pixel in the image to obtain the gradient components in the horizontal and vertical directions.

[0095] S503 calculates the gradient amplitude of each pixel;

[0096] S504 calculates the gradient direction of each pixel;

[0097] The gradient direction of each pixel is calculated using the following formula:

[0098] θ=arctan(Gx / Gy)

[0099] S505 divides the pixels into a plurality of direction groups according to the gradient direction;

[0100] The pixels are divided into several direction groups according to the gradient direction, such as 0°~22.5°, 22.5°~67.5°, 67.5°~112.5°, etc.

[0101] S506: for each pixel point, check whether its gradient amplitude is greater than the gradient amplitudes of the pixels along the gradient direction in its two neighborhoods;

[0102] The local maxima are retained as edge points and the rest are set to zero.

[0103] S507 sets a high threshold and a low threshold. Pixels above the high threshold are considered strong edge points, and pixels below the low threshold are considered background.

[0104] Set two thresholds, a high threshold and a low threshold. Points above the high threshold are considered strong edge points, and points below the low threshold are marked as background. Points between the two are considered edge points if they are connected to strong edge points.

[0105] S404: Screen out the contour that best matches the garment contour according to the characteristic factors of the contour, wherein the characteristic factors include area, perimeter, center point, minimum circumscribed rectangle, and minimum circumscribed circle.

[0106] The specific steps include:

[0107] S601, for each contour, calculate contour characteristics;

[0108] Area calculation: Use the area of ​​a contour to assess the size of the contour.

[0109] Perimeter calculation: Calculate the perimeter of a contour, which is used to describe the length of the contour.

[0110] Center Point Calculation: Calculates the centroid or center point of a contour.

[0111] Minimum bounding rectangle calculation: Find the smallest rectangle that can completely contain the outline.

[0112] Minimum circumscribed circle calculation: Find the smallest circle that can completely contain the contour.

[0113] S602 sets a matching threshold range to exclude contours outside the range.

[0114] According to the actual application scenario, a reasonable threshold range is set, such as the threshold of area, perimeter, etc., to exclude contours that do not meet the conditions. According to the set threshold range, the contour that best matches the clothing contour is screened out.

[0115] S103 converts the clothing outline into vector graphics data;

[0116] The specific steps include:

[0117] S701 uses OpenCV to extract contours and convert them into vector paths;

[0118] The specific steps include:

[0119] S801 represents each contour point with coordinates;

[0120] Coordinate representation: Convert each point in the contour into an (x, y) pair in a two-dimensional coordinate system, where x represents the horizontal position and y represents the vertical position.

[0121] Contour point extraction: Use the findContours function in OpenCV to extract contour points and store them as a point set.

[0122] S802 calculates the distance between adjacent points and sorts them based on the distance;

[0123] Calculate the Euclidean distance between each pair of adjacent points on the contour. Sort the adjacent points according to their distance, which helps in the subsequent point simplification and curve fitting process.

[0124] S803 uses the Douglas-Peucker algorithm to reduce the number of contour points;

[0125] Apply the Douglas-Peucker algorithm to reduce the number of contour points while maintaining the essential shape characteristics of the contour. Set a tolerance value that determines the maximum allowable deviation between the simplified contour and the original contour.

[0126] S804 uses a cubic Bezier curve to fit the contour points to create a smooth curve segment to obtain a vector path.

[0127] Based on the simplified contour points, a cubic Bezier curve is used for fitting to generate smooth curve segments. Appropriate control points are calculated for each curve segment to ensure that the curve is smooth and conforms to the actual shape of the contour. All curve segments are connected to form a continuous vector path.

[0128] S702 saves the vector path as an SVG file.

[0129] Use Scalable Vector Graphics (SVG) file format to save the vector path. Convert the vector path obtained above to SVG file format and save it to disk. Ensure that the generated SVG file can be displayed correctly on multiple devices and platforms

[0130] S104 generates a digital model of clothing according to the vector graphic data.

[0131] The vector graphics data is read from the SVG file generated in the previous step (S103).

[0132] Parse the SVG file to get the coordinates of the outline points and any additional attributes such as color or material information.

[0133] Decide which 3D modeling software or development environment to use, such as Blender, Maya, or specialized clothing design software such as Marvelous Designer.

[0134] Create a base mesh or surface in 3D space based on the vector paths in the SVG file. Subdivide the base mesh to increase detail and precision. Define the material properties used for the garment, such as cotton, silk, etc., and simulate the physical properties of these materials in the software. Use texture maps to add details to the model, such as patterns, textures, or colors. Adjust the shape of the 3D model according to the requirements of the garment design to ensure that it meets the design specifications. Add additional details such as pleats, seams, buttons or other decorative elements. If the garment is made up of multiple parts, you need to assemble the parts together. Add animation effects to the garment model, such as simulating the movement when wearing it, so as to see how it looks in different postures.

[0135] Second embodiment

[0136] The invention provides an image vectorized clothing entry system, comprising a clothing scanning module, a contour recognition module, a contour conversion module and a model generation module; the clothing scanning module is used for scanning clothing to obtain a clothing digital image; the contour recognition module is used for identifying the clothing contour in the clothing digital image by using an image processing technology; the contour conversion module is used for converting the clothing contour into vector graphic data; and the model generation module is used for generating a digital model of clothing according to the vector graphic data.

[0137] In this embodiment, the system is an integrated solution that aims to transform physical clothing into highly accurate digital models through a series of automated processes. The entire system includes four main modules: clothing scanning module, contour recognition module, contour conversion module and model generation module.

[0138] The clothing scanning module scans the physical clothing with high precision and obtains the digital image of the clothing. The physical clothing to be scanned is placed in the scanning area. A high-resolution camera or a multi-angle scanner is used to capture all sides of the clothing. The captured image is pre-processed, such as denoising, brightness and contrast adjustment, to improve the accuracy of subsequent processing.

[0139] The contour recognition module uses edge detection algorithms (such as Canny edge detection) to extract edge information from the digital image of clothing. Image segmentation methods (such as threshold segmentation or machine learning algorithms) are used to distinguish clothing from the background. The extracted edge information is analyzed to identify the main contour lines of the clothing. Some post-processing is also required, such as removing noise points and smoothing contour lines.

[0140] Contour conversion module: Use vectorization technology to convert the pixel-level contours output by the contour recognition module into mathematically described curves, such as Bezier curves or spline curves. Curve fitting is required during the conversion process to ensure that the generated vector graphics not only maintains the details of the original image but also has good editability. Finally, SVG (scalable vector graphics) or other vector format files are output for subsequent use.

[0141] The model generation module imports the vector graphics data generated by the outline conversion module. The 2D outline data is expanded into a model in 3D space using 3D modeling software. Details such as textures and material properties are added to make the model more realistic. Physical simulation and animation of the model are performed to enable it to show real dynamic effects in a virtual environment.

[0142] Overview of the overall workflow: Use high-precision scanning equipment to obtain digital images of physical clothing. Use image processing technology to identify the outline of the clothing from the digital image. Convert the outline into vector graphics data for subsequent processing. Build a 3D clothing model based on the vector graphics data, and add necessary details and simulate.

[0143] Through such a process, the system can efficiently transform physical clothing into high-quality digital models, which can be used in various application scenarios such as e-commerce and virtual try-on.

[0144] What is disclosed above is only a preferred embodiment of the present invention, and certainly cannot be used to limit the scope of rights of the present invention. Ordinary technicians in this field understand that all or part of the processes of the above embodiment and equivalent changes made according to the claims of the present invention still fall within the scope of the invention.

Claims

1. A method for image vectorization clothing input, It is characterized in that The method comprises: scanning the clothing to obtain a digital image of the clothing; Using image processing technology to identify the garment contour in the garment digital image; Converting the garment outline into vector graphics data; A digital model of the garment is generated according to the vector graphic data.

2. The method for entering clothing into image vectorization according to claim 1, characterized in that: The specific steps of scanning the clothing to obtain the digital image of the clothing include: Iron the garment flat before scanning to ensure there are no wrinkles; Adjust the scanner to acquire garment images according to the resolution and color mode of the desired images; The clothing images are stitched into a complete clothing digital image.

3. The method for entering clothing into image vectorization according to claim 2, characterized in that: The specific steps of stitching the clothing images into a complete clothing digital image include: Adjust the curve smoothness of the clothing image and remove the overlapping parts; The clothing image is segmented and adjusted according to the required size and shape to obtain clothing units; All clothing units are spliced ​​together to obtain a clothing digital image.

4. The method for entering clothing into image vectorization according to claim 3, characterized in that: The specific steps of using image processing technology to identify the clothing outline in the clothing digital image include: Converting digital images of clothing into grayscale images; Apply filters to remove noise from grayscale images; Use the Sobel operator to calculate the gradient of the grayscale image and find the edge of the image; The contour that best matches the garment contour is selected according to the characteristic factors of the contour, wherein the characteristic factors include area, perimeter, center point, minimum circumscribed rectangle and minimum circumscribed circle.

5. The method for entering clothing into image vectorization according to claim 4, characterized in that: The specific steps of using the Sobel operator to calculate the gradient of the grayscale image and find the edge of the image include: Normalize the pixel value range to -1 to 1; Use the two matrices of the Sobel operator to perform convolution operations with each pixel in the image to obtain the gradient components in the horizontal and vertical directions; Calculate the gradient magnitude of each pixel; Calculate the gradient direction of each pixel; Divide pixels into multiple direction groups according to the gradient direction; For each pixel, check whether its gradient magnitude is greater than the gradient magnitudes of the pixels along the gradient direction in its two neighborhoods; Set a high threshold and a low threshold. Pixels above the high threshold are considered strong edge points; pixels below the low threshold are considered background.

6. The method for entering clothing into image vectorization according to claim 5, characterized in that: The specific steps of selecting the contour that best matches the garment contour according to the characteristic factors of the contour include: For each contour, contour characteristics are calculated; Set a matching threshold range to exclude contours outside the range.

7. The method for entering clothing into image vectorization according to claim 6, characterized in that: The specific steps of converting the garment outline into vector graphics data include: Use OpenCV to extract contours and convert them into vector paths; Save vector paths as SVG files.

8. The method for entering clothing into image vectorization according to claim 7, characterized in that: The specific steps of using OpenCV to extract the contour and convert it into a vector path include: Represent each contour point with coordinates; Calculate the distance between adjacent points and sort them based on the distance; Use the Douglas-Peucker algorithm to reduce the number of contour points; The contour points are fitted using cubic Bezier curves to create smooth curve segments to obtain vector paths.

9. An image vectorized clothing entry system, characterized in that: It includes a clothing scanning module, a contour recognition module, a contour conversion module and a model generation module; The clothing scanning module is used to scan the clothing to obtain a digital image of the clothing; The outline recognition module is used to recognize the clothing outline in the clothing digital image using image processing technology; The outline conversion module is used to convert the clothing outline into vector graphics data; The model generation module is used to generate a digital model of clothing according to the vector graphics data.

Images