Method for reducing a digital image to yarn colors determİned by the designer and converting it into a knitting code
By enabling designers to specify yarn colors and algorithms, the method ensures accurate and detailed image transfer to knitting machines, enhancing design flexibility and sustainability.
Patent Information
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- FIRAT UNIVSI REKTORLUGU
- Filing Date
- 2024-12-18
- Publication Date
- 2026-05-28
AI Technical Summary
Existing methods for transferring digital images to knitting machines face challenges such as loss of detail, lack of color accuracy, and limited control over yarn selection, leading to deviations from the original design and reduced aesthetic quality.
A method that allows designers to directly specify yarn colors and select color reduction algorithms, using automated evaluation metrics to ensure color accuracy and detail preservation, and supports flexible design options.
Enables high-quality transfer of digital images to jacquard knitted fabrics by maintaining original design details and colors, reducing manual intervention, and promoting sustainable yarn use.
Smart Images
Figure 00000010_0000 
Figure 00000011_0000
Abstract
Description
[0001] DESCRIPTION
[0002] METHOD FOR REDUCING A DIGITAL IMAGE TO YARN COLORS DETERMINED BY THE DESIGNER AND CONVERTING IT INTO A KNITTING CODE
[0003] TECHNICAL FIELD
[0004] The invention relates to a method encompassing processes for making digital images suitable for knitting machines. Our invention resides at the intersection of digital image processing and textile technologies. Existing applications in this field include the process of optimizing digital images to align with the constraints of knitting machines, which are required to operate with a limited number of yam colors.
[0005] PRIOR ART
[0006] The process of transferring digital images to textile surfaces often faces challenges such as loss of detail, lack of color accuracy, and resolution limitations. Traditional methods used in the textile industry frequently require manual intervention, making it difficult to optimize images to fully preserve their details, colors, and textures.
[0007] It enables the processing of digital "adversarial" (misleading) images onto jacquard knitted fabrics. This method uses a combination of image processing techniques and knitting machines to optimize the color and resolution capacity of the image to fit the knitting machine and transfers the image onto the knitted fabric. However, in this process, the image must be reduced according to the machine’s limited color capacity (usually 8 colors).
[0008] The method operates as follows:
[0009] • The digital image is processed in high resolution, and the number of colors is reduced — e.g., to 8 colors — using specific algorithms to match the color limitations of the knitting machine.
[0010] • For these 8 colors in the reduced image, the closest yarn colors are determined, or yams are blended to achieve the closest possible match. • These yarn colors are placed onto the yam carriers of the knitting machine, and then the image is processed and transferred onto the fabric.
[0011] 1. Lack of Flexibility in Color Selection by the Designer
[0012] In this method in the known art, the designer cannot directly determine the yam colors to be used. The image is automatically reduced — e.g., to eight colors — to make it suitable for the machine, and then the closest yam colors to these eight colors are automatically selected. In this case, the designer cannot adjust the colors in the image to match the available yam colors.
[0013] The lack of freedom to directly specify yam colors causes the image to lose its original hues and subtleties. Since the eight-color palette is chosen based on the machine’s limited yam colors, it becomes difficult to accurately reflect the original design colors.
[0014] 2. Limited Control Over Yarn Selection Leads to Loss of Detail
[0015] In this method in the known art, an automatic color selection is made according to the eight-color capacity of the machine during image reduction, but the designer has no opportunity to intervene at this stage. The inability to directly specify yam colors causes significant loss of detail, especially for complex and multi-colored designs.
[0016] Without yam colors specified by the designer, this automatic reduction process dulls the details of the image. Original patterns are either entirely lost or transferred to the fabric in a more faded and irregular manner. Unnatural tonal shifts and streaks occur in color transitions, diminishing the product’s aesthetic quality.
[0017] 3. Lack of Options to Ensure Color Accuracy and Compatibility
[0018] In this method in the known art, the color reduction of the image is carried out using a single algorithm, which automatically operates to match, for example, the eight colors that the knitting machine can support. However, the designer is not provided with options to choose between different algorithms or to manually adjust the color accuracy of the reduced image.
[0019] This limitation prevents the designer from optimizing the image’s detail and color accuracy. The inability to select from different color reduction algorithms makes it difficult to preserve accurate color transitions in complex designs. As a result, the product deviates from the original design and loses realism, as the designer cannot accurately translate the details onto the fabric.
[0020] Bu smirlama, tasarimcimn gdrselin detay ve renk dogrulugunu optimize etmesini engeller. Farkli renk indirgeme algoritmalan arasmda segim yapamamak, karmagik tasanmlann dogru renk gegiglerini korumasim zorlagtirir. Sonug olarak, tasanmci detaylan dogru bir gekilde kumaga yansitamadigi igin urun orijinal tasanmdan uzaklagir ve gergekgilik kaybi yagamr. Absence of Automated Metrics for Quality and Color Accuracy Evaluation
[0021] In this method in the known art, no automated metrics are used to evaluate the reduced image for quality and color accuracy. Any quality loss, color distortion, or detail degradation occurring during the reduction process must be assessed manually.
[0022] The absence of automated evaluation metrics makes it difficult for designers to detect quality loss at the outset. Manual assessments, especially of color accuracy, have a high margin of error. This results in uncontrollable quality problems and aesthetic defects during the fabric transfer process. Inability to Select Yarns Closely Matching the Original Color Palette
[0023] In this method in the known art, after the image is reduced — e.g., to eight colors — the closest yam colors are placed on the machine, but the yams are not selected according to the original color palette. In this case, the reduced colors may not accurately correspond to the original colors, and the primary colors of the image may be lost.
[0024] The designer fails to accurately transfer the original color palette from the image to the fabric, causing the design to lose its original impact. For instance, subtle tonal differences present in the original design may not be adequately represented by the yam colors, reducing the visual impact of the pattern. Due to the technical issues described in detail above, it is not possible to fully reflect the details of a digital "adversarial" image onto jacquard knitted fabric..
[0025] BRIEF DESCRIPTION OF THE INVENTION
[0026] The invention offers an innovative solution to technical problems such as color accuracy, detail loss, and the need for manual intervention during the process of transferring digital images onto textile surfaces with high accuracy and quality. In the textile industry, the necessity of reducing digital images to yam colors that can be used by knitting machines prevents the preservation of original image details and the accuracy of colors. Traditional methods are limited by the color capacity of knitting machines, and designers cannot exert direct control over yarn colors. This results in the loss of image details and a deviation from the desired quality standards.
[0027] DESCRIPTION OF THE FIGURES
[0028] Figure 1. Flow Diagram of the Invention
[0029] Figure 2. System Architecture of the Interface and Background Modules for Knitting Design
[0030] DETAILED DESCRIPTION OF THE INVENTION
[0031] Our invention enables digital images to be optimized for knitting machines by allowing the designer to directly specify the yam colors. According to the yam colors chosen by the designer, the image is processed using the most appropriate color reduction algorithms. In this process, each pixel is saved by aligning it with knitting codes according to the RGB color model value of the yam color specified by the designer. Thus, a high-quality knitting pattern is obtained while maintaining color accuracy in the image.
[0032] With the method subject to the invention, the designer can preserve the color accuracy of the image by directly determining their own yam colors. In this way, digital images are brought closer to the original design, meeting the need for color accuracy and quality. The color reduction algorithms of the image can be selected by the designer through the method subject to the invention, and the most suitable algorithm is used to minimize detail loss. The image is automatically evaluated for quality and color accuracy, and the best result is presented. Since the designer can directly determine the yam colors, they can create more original and detailed designs. This feature will make a significant contribution to the fashion and textile industry, which requires the production of unique designs.
[0033] In traditional methods, processes such as color adaptation, posterization, and manual color selection are time-consuming and impose a heavy workload on the designer. Since the method subject to our invention automatically performs color selection and image resizing, it significantly reduces the need for manual processing. Presenting alternative results generated by different algorithms to the designer allows them to easily select the best result without manual intervention. Thus, while maintaining quality in the design process, the designer is provided with significant time savings.
[0034] Our invention gives the designer the freedom to determine yam colors. The designer optimizes the image according to the specified yam colors, thereby ensuring color accuracy. This feature ensures the preservation of the original design’s color accuracy and enables the image to be directly aligned with the yam colors.
[0035] By selecting between different color reduction algorithms, the designer minimizes detail loss. This provides a critical advantage, especially for complex and multi-colored designs. The ability to choose between different algorithms allows the designer to achieve a more realistic result while preserving fine details and nuances.
[0036] Our invention includes evaluation metrics that automatically measure the quality and color accuracy of the image. With quality metrics such as CIEDE2000, SSIM, MSE, and PSNR, the versions of the image processed with different algorithms are evaluated, and the best result is presented automatically. In this way, the error rate in the manual quality control process is reduced.
[0037] By aligning the image with yam colors and incorporating automated quality evaluation processes, the need for manual handling is reduced. This user-friendly approach speeds up the process and increases work efficiency by reducing the error rate.
[0038] Our invention allows designers to create more flexible and unique designs as they can use their custom yam color palettes and make algorithm selections. This flexibility increases diversity in the textile industry while preserving high- resolution image quality.
[0039] Since our invention enables the utilization of stock yams, it allows for the use of idle yams, thereby providing a cost-effective and environmentally friendly production process. By reducing the need for unnecessary yam procurement, it facilitates more sustainable production and offers an eco-friendly alternative in the textile industry..
[0040] Application Steps of the Method Subject to Our Invention;
[0041] - The designer uploads the digital image to the application (Image Upload).
[0042] - The designer selects the yam colors (or the RGB values of the colors) to be used in the image through the application (Selecting Yam Colors).
[0043] - The capacity of the knitting machine to be used (1 pixel per needle) is entered into the application by the designer (Entering Machine Capacity).
[0044] - The application automatically resizes the image to the capacity of the knitting machine and checks for compatibility with the machine capacity (Automatic Resizing of Image and Comparison with Machine Capacity).
[0045] - The colors of the image are reduced according to the yam colors specified by the designer using Floyd-Steinberg error diffusion, adaptive color reduction, color and brightness optimization, and block-based color reduction algorithms (Applying Color Reduction Algorithms).
[0046] - The application presents the designer with a preview of the effect of each algorithm on the image, allowing the designer to select the appropriate algorithm based on these preview options (Previewing Algorithms).
[0047] - Each reduced image is evaluated using various quality and color accuracy metrics such as CIEDE2000, SSIM (Structural Similarity Index), MSE (Mean Squared Error), PSNR (Peak Signal-to-Noise Ratio), and VIF (Visual Information Fidelity) (Evaluating Color Accuracy and Quality).
[0048] - The image with the best results is automatically identified based on quality and color accuracy metrics (Selecting the Best Quality Image).
[0049] - The application presents other algorithm options to the designer along with the best result, allowing the designer to make the final selection (Presenting Options to the Designer).
[0050] - The image selected by the designer is recreated in BMP format that the knitting machine can use, along with the knitting codes specified by the designer. In this step, each RGB color value is converted according to the knitting code number specified in the knitting design software. The image is saved according to the knitting codes so that each color corresponds to a specific knitting code, saved as a BMP file. This ensures that the correct knitting pattern is processed when the design is transferred to the machine (Creating the Selected Image in BMP Format).
[0051] - The BMP format image is uploaded to the knitting design software (Upload to Knitting Design Software).
[0052] - The yams corresponding to each color in the image are attached to the relevant yam carrier of the knitting machine (Attaching the Yams to the Yam Carrier).
[0053] - The machine is operated, and the visual design is transformed into a product, such as knitwear or fabric, through the knitting process (Pattern to Product).
Claims
CLAIMS1. A method for reducing the digital image to the yam colors determined by the designer and converting it into knitting codes, characterized by;- The designer uploads the digital image to the application (Image Upload).- The designer selects the yarn colors (or the RGB values of the colors) to be used in the image through the application (Selecting Yam Colors).- The capacity of the knitting machine to be used (1 pixel per needle) is entered into the application by the designer (Entering Machine Capacity).- The application automatically resizes the image to the capacity of the knitting machine and checks for compatibility with the machine capacity (Automatic Resizing of Image and Comparison with Machine Capacity).- The colors of the image are reduced according to the yam colors specified by the designer using Floyd-Steinberg error diffusion, adaptive color reduction, color and brightness optimization, and block-based color reduction algorithms (Applying Color Reduction Algorithms).- The application presents the designer with a preview of the effect of each algorithm on the image, allowing the designer to select the appropriate algorithm based on these preview options (Previewing Algorithms).- Each reduced image is evaluated using various quality and color accuracy metrics such as CIEDE2000, SSIM (Structural Similarity Index), MSE (Mean Squared Error), PSNR (Peak Signal-to-Noise Ratio), and VIF (Visual Information Fidelity) (Evaluating Color Accuracy and Quality).- The image with the best results is automatically identified based on quality and color accuracy metrics (Selecting the Best Quality Image).- The application presents other algorithm options to the designer along with the best result, allowing the designer to make the final selection (Presenting Options to the Designer).- The image selected by the designer is recreated in BMP format that the knitting machine can use, along with the knitting codes specified by the designer. In this step, each RGB color value is converted according to the knitting code number specified in the knitting design software. The image is saved according to the knitting codes so that each color corresponds to a specific knitting code, saved as a BMP file. This ensures that the correct knitting pattern is processed when the design is transferred to the machine (Creating the Selected Image in BMP Format).- The BMP format image is uploaded to the knitting design software (Upload to Knitting Design Software).- The yarns corresponding to each color in the image are attached to the relevant yam carrier of the knitting machine (Attaching the Yarns to the Yam Carrier).- The machine is operated, and the visual design is transformed into a product, such as knitwear or fabric, through the knitting process (Pattern to Product).