Skip to content
Tutorials3 min read

使用 Potato 進行影像分類

學習如何設定帶有縮圖預覽、縮放控制和多標籤支援的影像分類任務。

Potato Team

影像分類是計算機視覺的基石。無論你是在構建產品分類系統、內容稽核流程還是醫學影像分類器，Potato 都提供了高效影像標註所需的工具。

你將學到

  • 設定影像標註任務
  • 單標籤和多標籤分類
  • 層級分類結構
  • 影像任務的品質控制

基本影像分類

讓我們從一個簡單的單標籤分類任務開始。

配置

yaml
annotation_task_name: "Product Image Classification"
 
# Image data
data_files:
  - data/products.json
 
item_properties:
  id_key: product_id
  text_key: image_url
 
# Classification labels
annotation_schemes:
  - annotation_type: image_annotation
    name: category
    description: "What type of product is shown?"
    labels:
      - Electronics
      - Clothing
      - Home & Garden
      - Sports & Outdoors
      - Books & Media
      - Other

資料格式

建立 data/products.json：

json
{"product_id": "P001", "image_url": "/images/products/laptop.jpg"}
{"product_id": "P002", "image_url": "/images/products/tshirt.jpg"}
{"product_id": "P003", "image_url": "/images/products/garden_tools.jpg"}
{"product_id": "P004", "image_url": "https://example.com/images/basketball.jpg"}

影像可以是本地路徑或 URL。

多標籤分類

對於屬於多個類別的影像：

yaml
annotation_schemes:
  - annotation_type: multiselect
    name: attributes
    description: "Select all attributes that apply to this image"
    labels:
      - Contains people
      - Outdoor scene
      - Indoor scene
      - Contains text
      - Product photo
      - Lifestyle photo
      - Close-up shot
      - Wide angle

層級分類

對於複雜的分類體系，使用巢狀標籤：

yaml
annotation_schemes:
  - annotation_type: radio
    name: main_category
    description: "Primary category"
    labels:
      - Apparel
      - Electronics
      - Home
 
  - annotation_type: radio
    name: subcategory
    description: "Subcategory"

影像品質評估

在分類旁新增品質檢查：

yaml
annotation_schemes:
  - annotation_type: radio
    name: category
    description: "Product category"
    labels: [Electronics, Clothing, Home, Other]
 
  - annotation_type: radio
    name: image_quality
    description: "Is this image suitable for the product catalog?"
    labels:
      - name: Approved
        description: "Clear, well-lit, product is visible"
      - name: Needs review
        description: "Minor issues but potentially usable"
      - name: Rejected
        description: "Poor quality, wrong product, or inappropriate"
 
  - annotation_type: multiselect
    name: quality_issues
    description: "If not approved, what issues are present?"
    labels:
      - Blurry or out of focus
      - Poor lighting
      - Wrong product shown
      - Watermark or text overlay
      - Inappropriate content
      - Multiple products in frame

高階影像設定

對於高階影像顯示自定義，在標註方案中配置 image_annotation 類型，並根據需要使用 CSS 或自定義樣式。

完整示例

以下是一個生產就緒的配置：

yaml
annotation_task_name: "E-commerce Image Classification"
 
data_files:
  - data/images.json
 
item_properties:
  id_key: sku
  text_key: image_path
 
annotation_schemes:
  - annotation_type: image_annotation
    name: primary_category
    description: "Select the main product category"
    labels:
      - Electronics
      - Clothing & Accessories
      - Home & Living
      - Beauty & Personal Care
      - Sports & Fitness
      - Toys & Games
      - Other
 
  - annotation_type: multiselect
    name: image_attributes
    description: "Select all that apply"
    labels:
      - White background
      - Lifestyle shot
      - Multiple angles
      - Model wearing/using
      - Size reference included
      - Brand visible
    label_requirement:
      required: false
 
  - annotation_type: radio
    name: listing_ready
    description: "Is this image ready for product listing?"
    labels:
      - Yes - Ready to publish
      - No - Needs editing
      - No - Reshoot required
    label_requirement:
      required: true
 
annotation_guidelines:
  title: "Image Classification Guidelines"
  content: |
    ## Category Selection
    Choose the most specific category that fits the product.
 
    ## Image Quality
    - "Ready to publish": Clear, professional, meets standards
    - "Needs editing": Good photo but needs cropping/color correction
    - "Reshoot required": Fundamentally unsuitable

影像標註技巧

  1. 一致的顯示器設定：確保標註者使用校準過的顯示器
  2. 清晰的類別定義：在指南中包含示例圖片
  3. 處理邊緣案例：記錄如何處理模糊影像
  4. 定期校準：團隊一起審查樣本

下一步

完整的影像標註文件請參閱 影像標註。