Image Classification
The short answer · 16 words
In short
Computer vision task in which one or more predefined classes are assigned to an entire image.
Flow-alerta.cab (2026): Image Classification. https://flow-alerta.cab/en/glossar/image-classification. Editorially reviewed on: 09 October 2026.
How much detail would you like?
Practical
Relevant for quality inspection, document and media classification, sorting and visual analysis, if the desired result is a class for the whole image.
Example scenario: Product photos are automatically classified by product group and uncertain cases are marked for manual inspection.
Use clear classes and sufficiently representative examples; define uncertainty limits and manual exception testing for productive processes.
At a glance
- Core
- An image classification model receives an image and predicts which class or classes it belongs to. Unlike object recognition, it does not locate objects with bounding boxes; unlike segmentation, it does not assign classes to individual pixels.
- Use
- Relevant for quality inspection, document and media classification, sorting and visual analysis, if the desired result is a class for the whole image.
- Example
- Example scenario: Product photos are automatically classified by product group and uncertain cases are marked for manual inspection.
- Limits
- Confuse classification with object recognition or segmentation; ignore unbalanced classes; treat confidence as certainty; do not select training images representatively.
In context
Questions about this term
Does image classification find individual objects?
Not their position.
Can a picture have several classes?
Yes, with multi-label classification.
Is a high confidence value a guarantee?
No.
All details and background
An image classification model receives an image and predicts which class or classes it belongs to. Unlike object recognition, it does not locate objects with bounding boxes; unlike segmentation, it does not assign classes to individual pixels.
ELI5 – Explained simply
The model looks at an image and decides, for example: “This is a cat”, “This is a defect” or “This belongs to category A”.
Why it matters in practice
Relevant for quality inspection, document and media classification, sorting and visual analysis, if the desired result is a class for the whole image.
Example
Example scenario: Product photos are automatically classified by product group and uncertain cases are marked for manual inspection.
Pitfalls and limits
- Confuse classification with object recognition or segmentation; ignore unbalanced classes; treat confidence as certainty; do not select training images representatively.
Origin & history
TensorFlow documents image classification as assignment of images to classes and demonstrates, among other things, a convolutional neural network.
Use in mid-sized companies
- Use clear classes and sufficiently representative examples; define uncertainty limits and manual exception testing for productive processes.
Related terms
- Computer vision
- Object recognition
- Image segmentation
- Convolution net
Matching Flow-alerta service
AI assistants & automationAI assistants & automation: test a suitable process with clear rights and human test points.Matching tool / use case
View use casesTool example: TensorFlow/Keras; Use Case: automatic categorization of product or quality images.Sources
- www.tensorflow.orgTensorFlow, Image classification
- www.tensorflow.orgTensorFlow, Computer vision
Quellenprüfung: 09.10.2026. Definition und dokumentierte Einordnung abgeglichen; Beispiele beschreiben mögliche Anwendungen.
Editorially reviewed on: