# US 10,769,784 B2 · Image analyzing method and electrical device

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

- Number: US 10,769,784 B2
- Document ID: US10769784B2
- Status: Granted
- Kind: B2
- Application number: 16/228,778
- Filed: 2018-12-21
- Published: 2020-06-25
- Granted: 2020-09-08
- Earliest claimed: 2018-12-21
- Assignees: Metal Industries Research & Development Centre
- Applicants: Metal Industries Research & Development Centre
- Inventors: Sheng-Hong Yang; Jian-Jia Zeng; Bo-Wei Pan
- Web page: https://patentgrep.com/patent/US10769784B2

## CPC

- G06T 7/0012: {Biomedical image inspection}
- G06F 18/21: Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06N 3/045: Combinations of networks
- G06N 3/0464: Convolutional networks [CNN, ConvNet]
- G06N 3/08: Learning methods
- G06T 7/11: Region-based segmentation
- G06T 7/62: of area, perimeter, diameter or volume
- G06V 10/454: {Integrating the filters into a hierarchical structure, e.g. convolutional neural networks [CNN]}
- G06V 10/50: by performing operations within image blocks; by using histograms, e.g. histogram of oriented gradients [HoG]; by summing image-intensity values; Projection analysis
- G06V 10/82: using neural networks
- G06V 20/695: {Preprocessing, e.g. image segmentation}
- G06N 3/048: Activation functions
- G06N 20/10: using kernel methods, e.g. support vector machines [SVM]
- G06T 2207/20081: Training; Learning
- G06T 2207/20084: Artificial neural networks [ANN]
- G06T 2207/30024: Cell structures in vitro; Tissue sections in vitro
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING

## Abstract

An image analyzing method is provided and includes: extracting a first feature vector according to global information of a digital image; dividing the digital image into multiple regions, and inputting each region into a convolutional neural network to obtain a second feature vector; merging the first feature vector with the second feature vectors to obtain a third feature vector; and performing an image analyzing process according to the third feature vector.

## Claims

### Claim 1

An image analyzing method for an electrical device, wherein the image analyzing method comprises:
extracting a first feature vector according to global information of a digital image, wherein the digital image is a medical image, and the first feature vector consists of a nuclear area ratio, an average nuclear brightness, an average cytoplasmic brightness, and a nuclear and cytoplasmic brightness ratio;
dividing the digital image into a plurality of regions, and inputting each of the regions into a convolutional neural network to obtain a plurality of feature maps;
applying a global filter to the feature maps to obtain a plurality of feature values, wherein a size of the global filter is identical to a size of each of the feature maps, and the feature values corresponding to one of the regions constitute one of a plurality of second feature vectors;
merging the first feature vector with the second feature vectors to obtain a third feature vector; and
performing an image analyzing process according to the third feature vector.

### Claim 2 (depends on 1)

The image analyzing method of claim 1, wherein the nuclear area ratio is a ratio of an area of nuclei to an area of cells; the average nuclear brightness is an average of gray levels of the nuclei; the average cytoplasmic brightness is an average of gray levels of cytoplasm; and the nuclear and cytoplasmic brightness ratio is a ratio of the average nuclear brightness to the average cytoplasmic brightness.

### Claim 3 (depends on 1)

The image analyzing method of claim 1, wherein the regions are not overlapped with each other.

### Claim 4

An electrical device, comprising:
a memory, storing a plurality of instructions; and
a processor, executing the instructions to perform steps of:
extracting a first feature vector according to global information of a digital image, wherein the digital image is a medical image, and the first feature vector consists of a nuclear area ratio, an average nuclear brightness, an average cytoplasmic brightness, and a nuclear and cytoplasmic brightness ratio;
dividing the digital image into a plurality of regions, and inputting each of the regions into a convolutional neural network to obtain a plurality of feature maps;
applying a global filter to the feature maps to obtain a plurality of feature values, wherein a size of the global filter is identical to a size of each of the feature maps, and the feature values corresponding to one of the regions constitute one of a plurality of second feature vectors;
merging the first feature vector with the second feature vectors to obtain a third feature vector; and
performing an image analyzing process according to the third feature vector.