CS5720 - Week 4
Slide 74 of 80
Pooling Layer Benefits
Pooling Operations
What is Pooling?
Pooling is a
downsampling operation
that reduces the spatial dimensions of feature maps while preserving important information.
MAX
Max Pooling
Takes maximum value
AVG
Average Pooling
Computes average
GLB
Global Pooling
Entire feature map
ADP
Adaptive Pooling
Fixed output size
Key Benefits
🔽 Dimension Reduction
Reduces spatial dimensions, making computation more efficient and manageable.
🔄 Translation Invariance
Provides robustness to small translations and spatial variations in input.
🛡️ Reduces Overfitting
Acts as regularization by summarizing local features and reducing parameters.
👁️ Larger Receptive Field
Increases the effective receptive field of subsequent layers progressively.
⚡ Computational Efficiency
Reduces memory usage and computational load in deeper layers.
Interactive Pooling Demonstration
Input Feature Map (4×4)
→
Max Pooled Output (2×2)
Max Pooling
Average Pooling
New Input
Animate Process
Pooling Parameters:
Kernel Size: 2×2, Stride: 2, Padding: 0
Output Size = (Input - Kernel + 2×Padding) / Stride + 1
← Previous
Next →
Prepared by Dr. Gorkem Kar
Modal Title
×
Modal content goes here...