Detector24
Back to Model Catalogue
Alcoholic Beverages Detector
ImageRestricted Content

Alcoholic Beverages Detector

Detect alcoholic beverages in images including wine, beer, spirits, and cocktails. Enforce advertising compliance and content policies automatically.

Accuracy
96.1%
Avg. Speed
5.0s
Per Request
$0.0060
API Name
alcohol-detection

Bynn Alcoholic Beverages Detector

The Bynn Alcoholic Beverages Detector identifies and locates alcohol-related content in images, including bottles, glasses, cans, and people consuming alcoholic drinks. This model is essential for platforms enforcing age-restricted content policies and advertising regulations.

The Challenge

Alcohol advertising and content are heavily regulated worldwide to protect minors and public health. Platforms face legal obligations to prevent alcohol marketing from reaching underage audiences, while social media and user-generated content make enforcement increasingly complex.

The stakes are significant: regulatory fines for non-compliance, brand safety concerns for advertisers, and the societal impact of exposing young people to alcohol promotion. Manual review of millions of daily uploads is impossible, creating an urgent need for automated detection that can identify alcohol content across diverse contexts—from obvious beer advertisements to subtle influencer posts featuring cocktails.

For physical security and compliance, alcohol detection enables venue monitoring. CCTV systems can detect underage drinking in bars and clubs, alcohol consumption in prohibited areas, or policy violations at corporate events. Schools and youth facilities can monitor for contraband. Workplace safety systems can flag alcohol presence in restricted environments where impairment poses danger.

Model Overview

When provided with an image, the detector analyzes the visual content to identify alcoholic beverages and related items. The model uses advanced object detection with instance segmentation, providing not only the location of detected items but also pixel-level masks that precisely outline each detection.

Achieving 96.1% accuracy, the model can detect a wide variety of alcoholic beverages including beer, wine, spirits, cocktails, and people actively consuming or holding alcoholic drinks.

How It Works

The model employs open-vocabulary object detection technology, allowing it to recognize alcoholic beverages based on learned visual patterns. For each detected item, the system returns:

  • Bounding box coordinates: The rectangular region containing the detected object
  • Confidence score: A probability value (0.0-1.0) indicating detection certainty
  • Class name: The specific type of alcohol-related item detected
  • Segmentation mask: A pixel-level PNG mask outlining the exact shape of the detected object

Response Structure

The API returns a structured response containing:

  • num_detections: Total count of alcoholic items detected in the image
  • detections: Array of detection objects, each containing:
    • bbox: Bounding box coordinates {x1, y1, x2, y2}
    • score: Detection confidence (0.0-1.0)
    • class_name: Type of alcohol-related item
    • mask_png: Base64-encoded PNG segmentation mask
  • image_size: Original image dimensions {width, height}

Detected Classes

The model can identify the following alcohol-related items:

Beverages & Containers

  • Beer, beer bottle, beer can, beer mug
  • Wine, wine bottle, wine glass, champagne
  • Whiskey, vodka, rum, gin, tequila
  • Cocktail, martini, shot glass
  • Liquor, liquor bottle, keg
  • Generic alcohol, drink, beverage

People & Activities

  • Person drinking beer, person drinking wine
  • Person holding drink, person at bar
  • Drinking alcohol

Performance Metrics

Metric Value
Detection Accuracy 96.1%
Average Response Time 5000ms
Max File Size 20MB
Supported Formats GIF, JPEG, JPG, PNG, WebP

Use Cases

  • Age-Restricted Content Filtering: Automatically flag or restrict content showing alcohol for age-gated platforms
  • Advertising Compliance: Ensure alcohol advertisements are not displayed to underage audiences or in prohibited contexts
  • Social Media Moderation: Monitor user-generated content for alcohol-related posts that may violate community guidelines
  • E-commerce Verification: Verify product images for alcohol marketplace compliance
  • Brand Safety: Prevent brand advertisements from appearing alongside alcohol-related content
  • Influencer Content Review: Identify sponsored content featuring alcoholic beverages for disclosure requirements

Known Limitations

Important Considerations:

  • Partial Visibility: Heavily occluded or partially visible bottles may have lower detection confidence
  • Non-Alcoholic Lookalikes: Non-alcoholic beverages in similar containers (e.g., non-alcoholic beer) may be detected as alcoholic
  • Small Objects: Very small or distant alcohol containers may not be detected
  • Unusual Packaging: Craft beverages with non-standard packaging may have reduced detection rates
  • Context Limitations: The model detects visual presence, not context (e.g., cannot distinguish between drinking and holding for photography)

Disclaimers

This model provides probability scores, not definitive identification.

  • Screening Tool: Use as part of a broader content moderation strategy, not as the sole decision factor
  • Human Review: High-stakes decisions (account suspension, legal action) should involve human review
  • False Positives: Some non-alcoholic items may be flagged; adjust confidence thresholds based on your use case
  • Regional Variations: Alcohol regulations vary by jurisdiction; ensure your implementation aligns with local laws

Best Practice: Combine detection results with human review and contextual analysis for optimal content moderation outcomes.

API Reference

Version
2601
Jan 3, 2026
Avg. Processing
5.0s
Per Request
$0.006
Required Plan
trial

Input Parameters

Open-vocabulary object detection with segmentation. Detects objects using text prompts.

image_urlstringRequired

URL of the image for object detection

Example:
https://example.com/image.jpg

Response Fields

Object detection results with bounding boxes and segmentation masks

num_detectionsinteger

Number of objects detected

Example:
3
detectionsarray

Array of detected objects

Array Item Properties:
bboxobject

Bounding box {x1, y1, x2, y2}

{"x1":100,"y1":150,"x2":300,"y2":400}
scorefloat

Detection confidence (0.0-1.0)

0.95
class_idinteger

Class index

0
class_namestring

Detected object class name

person
mask_pngstring

Base64-encoded PNG segmentation mask

data:image/png;base64,...
image_sizeobject

Original image dimensions

Example:
{ "width": 1920, "height": 1080 }

Complete Example

Request

{
  "model": "alcohol-detection",
  "image_url": "https://example.com/image.jpg"
}

Response

{
  "success": true,
  "data": {
    "num_detections": 2,
    "detections": [
      {
        "bbox": {
          "x1": 100,
          "y1": 150,
          "x2": 300,
          "y2": 400
        },
        "score": 0.95,
        "class_id": 0,
        "class_name": "person",
        "mask_png": "data:image/png;base64,iVBORw0KGgo..."
      },
      {
        "bbox": {
          "x1": 400,
          "y1": 200,
          "x2": 600,
          "y2": 450
        },
        "score": 0.87,
        "class_id": 1,
        "class_name": "car",
        "mask_png": "data:image/png;base64,iVBORw0KGgo..."
      }
    ],
    "image_size": {
      "width": 1920,
      "height": 1080
    }
  }
}

Additional Information

Rate Limiting
If we throttle your request, you will receive a 429 HTTP error code along with an error message. You should then retry with an exponential back-off strategy, meaning that you should retry after 4 seconds, then 8 seconds, then 16 seconds, etc.
Supported Formats
gif, jpeg, jpg, png, webp
Maximum File Size
20MB
Tags:alcoholage-restrictedsegmentation

Ready to get started?

Integrate Alcoholic Beverages Detector into your application today with our easy-to-use API.

Learn more about image moderationBrowse all models