# US 2026/0087610 A1 · Method and Device for Predicting Product Manufacturing Index

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2026/0087610 A1
- Document ID: US20260087610A1
- Status: Published application
- Kind: A1
- Application number: 18/955,932
- Filed: 2024-11-21
- Published: 2026-03-26
- Earliest claimed: 2024-09-26
- Assignees: Wiwynn Corporation
- Applicants: Wiwynn Corporation
- Inventors: Ge-Wei Fang; Chih-Wei Chang
- Web page: https://patentgrep.com/patent/US20260087610A1

## CPC

- G06T 7/0004: {Industrial image inspection}
- G06V 10/25: Determination of region of interest [ROI] or a volume of interest [VOI]
- G06T 2207/20081: Training; Learning
- G06T 2207/30108: Industrial image inspection
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING

## Abstract

A method for predicting product manufacturing index includes performing an image preprocessing on a flat development drawing of a product to convert the flat development drawing into input data; performing a principal component analysis (PCA) on the input data to convert the input data into a principal component data; and using a first artificial intelligence (AI) model to predict a manufacturing index of the product according to the principal component data.

## Claims

### Claim 1

A method for predicting product manufacturing index, comprising:
Step P1) performing an image preprocessing on a flat development drawing of a product to convert the flat development drawing into input data;
Step P2) performing a principal component analysis (PCA) on the input data to convert the input data into a principal component data; and
Step P3) using a first artificial intelligence (AI) model to predict a manufacturing index of the product according to the principal component data.

### Claim 2 (depends on 1)

The method of claim 1, wherein the step P1) comprises:
Step P11) converting the flat development drawing into a binary image;
Step P12) cropping the binary image to retain a region of interest (ROI);
Step P13) normalizing a size of the ROI; and
Step P14) converting the normalized ROI into the input data represented as a one-dimensional array.

### Claim 3 (depends on 2)

The method of claim 2, wherein the step P11) comprises:
converting the flat development drawing into a grayscale image; and
performing binary conversion on the grayscale image to convert the grayscale image into the binary image.

### Claim 4 (depends on 1)

The method of claim 1, wherein the method comprises performing a model training process before executing the step P1), and the model training process comprises:
Step T1) performing the image preprocessing on each of a plurality of flat development drawings to convert the plurality of flat development drawings into a plurality of input data;
Step T2) performing the principal component analysis on each of the plurality of input data to convert the plurality of input data into a plurality of principal component data;
Step T3) determining a training dataset and a testing dataset from the plurality of principal component data;
Step T4) training a second initial model corresponding to a second AI model according to the training dataset and the testing dataset; and
Step T5) training a first initial model corresponding to the first AI model according to the training dataset and the testing dataset.

### Claim 5 (depends on 4)

The method of claim 4, wherein the step T4) comprises:
training the second initial model according to the training dataset and a training manufacturing index set corresponding to the training dataset; and
testing the second initial model according to the testing dataset and a testing manufacturing index set corresponding to the testing dataset.

### Claim 6 (depends on 5)

The method of claim 5, wherein the step T5) comprises:
training the first initial model according to the training dataset and the training manufacturing index set corresponding to the training dataset; and
testing the first initial model according to the testing dataset and the testing manufacturing index set corresponding to the testing dataset.

### Claim 7 (depends on 6)

The method of claim 6, wherein the model training process further comprises:
Step T6) when the first initial model and the second initial model do not pass the test, increasing sample completeness of the training dataset; and
returning to the step T4).

### Claim 8 (depends on 6)

The method of claim 6, wherein the model training process further comprises:
Step T7) when the second initial model passes the test, storing the second initial model as the second AI model; and
Step T8) when the first initial model does not pass the test and the second initial model passes the test, adjusting at least one hyperparameter of the first initial model and returning to the step T5).

### Claim 9 (depends on 8)

The method of claim 8, wherein the model training process further comprises:
Step T9) when the first initial model passes the test, storing the first initial model as the first AI model; and
Step T10) when the first initial model passes the test and no matter the second initial model passes the test or not, verifying a first accuracy of the first AI model and a second accuracy of the second AI model according to a verification testing set.

### Claim 10 (depends on 9)

The method of claim 9, wherein the model training process further comprises:
Step T11) when the first accuracy is lower than the second accuracy, returning to the step T8); or
Step T11) when the first accuracy is not lower than the second accuracy, finishing the model training process.

### Claim 11 (depends on 4)

The method of claim 4, wherein the second AI model is a linear regression model.

### Claim 12 (depends on 1)

The method of claim 1, wherein the first AI model is a support vector regression (SVR) model.

### Claim 13 (depends on 1)

The method of claim 1, wherein the product is a mechanical component, a printed circuit board, an interior space or a building, and the flat development drawing is a mechanical drawing, a printed circuit board layout, a panorama of interior design drawing or an architectural drawing.

### Claim 14 (depends on 1)

The method of claim 1, further comprising stitching multiple perspective views of the product into the flat development drawing before the step P1).

### Claim 15 (depends on 1)

The method of claim 1, wherein the manufacturing index is a total number of processes or a total number of molds.

### Claim 16 (depends on 1)

A first device for predicting product manufacturing index, comprising:
a first processing unit; and
a storing unit, coupled to the first processing unit, configured to store a program code, wherein the program code instructs the first processing unit to perform the method of claim 1.

### Claim 17 (depends on 16)

A second device, comprising:
an image capturing unit configured to capture a flat development drawing of a product; and
a communication unit coupled to the image capturing unit, configured to transmit the flat development drawing to the first device of claim 16, and receive a manufacturing index of the product from the first device.

### Claim 18 (depends on 17)

The second device of claim 17, further comprising:
a second processing unit coupled to the image capturing unit and the communication unit, and configured to stitch multiple perspective views of the product into the flat development drawing;
wherein the image capturing unit is configured to obtain the multiple perspective views of the product.
