Image Classification[AI]

## Image Classification[AI]

Automatically identify and classify images, outputting the category they belong to.

## Module Principle

Construct relevant inputs, outputs, and image displays for the module by inferring category IDs, confidence levels, etc., of recognition targets from trained models.

### Algorithm Workflow

Data Collection -> Annotation -> Preprocessing / Augmentation -> Model Selection -> Training -> Evaluation & Tuning -> NMS Post-processing -> Inference Deployment

### Algorithm Formula

1. Model Forward Propagation (Output Prediction Scores)
Given input image x, the deep network outputs category scores:
    z=f(x; θ)
    f: CNN/Transformer classification model
   θ: Model parameters (weights, biases)
   z∈R^c: Predicted scores for c categories
2. Softmax Normalization (Output Probabilities)
Convert scores to 0~1 probabilities that sum to 1:
3. Loss Function

## Usage Scenarios
Simple image classification

## Usage Method

### Pre/Post-Sequence Modules

Camera color image, video reading, display image, etc.

### Usage Example

![](/media/202603/2026-03-02_160405_3546690.9221680328166953.png)

## Parameter Configuration Description

![](/media/202603/2026-03-02_160449_6169290.6767479211079669.png)
## Output Description

![](/media/202603/2026-03-02_160505_5528900.5454441131950667.png)
## Function Attribute Description

![](/media/202603/2026-03-02_160532_0920720.23574778020882958.png)

---

Click here to jump to the corresponding Chinese manual page: [Image Classification[AI]](/doc/164)