Step by Step
C
Classification — one label for the whole image
Image classification answers "What is in this image?" with a single label describing the entire image, without identifying location.
Example: a classifier simply outputting the single label "cat" for an entire photo, without indicating where the cat is located within the frame.
D
Detection — what AND where, for every object
Object detection answers "What is where?" by drawing a bounding box around each individual object present and labeling each one separately.
Example: a detector identifying "cat" at the top-left of an image, "dog" at the bottom-right, and "person" in the center, each with its own separate bounding box and label.
L
Detection = classification + localization
Object detection can be understood as classification combined with localization — it doesn't just answer what's present, but also precisely where each instance is located.
Example: a single detection output combining both a class label ("dog") and a precise bounding box location for that specific dog within the image.
Y
YOLO — the famous real-time detector
YOLO (You Only Look Once) is a famous real-time object detector that processes the entire image in a single pass, making it fast enough for video applications.
Example: YOLO processing an entire video frame in one single forward pass to detect and label all objects present, achieving speeds fast enough for real-time video analysis.
Applied Walkthrough
1
A photo contains a cat, a dog, and a person, and a system needs to identify all three and roughly where each is located within the frame.
2
Ask: would image classification alone accomplish this? No — classification only outputs a single label for the whole image, and couldn't handle multiple distinct objects with separate locations.
3
Object detection is needed instead, since it can draw a separate bounding box around each of the three objects and label each one individually.
4
For real-time video applications, YOLO's single-pass processing approach makes it fast enough to perform this detection task at video frame rates, unlike slower two-stage detection approaches.
Exam Application
Exams test whether you understand the fundamental difference between classification (single whole-image label) and detection (per-object bounding box plus label), and whether you know YOLO as the famous example of a fast, single-pass, real-time-capable object detector.
⚠ Common Trap
The most common trap is assuming classification and detection are just different names for the same task. Classification only answers WHAT is in an image as a whole; detection answers WHAT and WHERE for potentially multiple distinct objects, requiring a fundamentally more complex output (multiple bounding boxes plus labels) than classification's single output label.
✓ Quick Self-Check
1. What question does image classification answer?
"What is in this image?" — with a single label for the whole image.
Tap to reveal / hide
2. What question does object detection answer?
"What is where?" — drawing bounding boxes around each object and labeling each one.
Tap to reveal / hide
3. How can object detection be understood in relation to classification?
As classification combined with localization.
Tap to reveal / hide
4. What makes YOLO notable among object detectors?
It processes the entire image in one single pass, making it fast enough for real-time video applications.
Tap to reveal / hide
5. Could image classification alone identify and locate three separate objects in one image?
No — classification only produces a single whole-image label, not separate per-object locations.
Tap to reveal / hide