Image Recognition Example

# Complete image recognition pipeline (conceptual walkthrough)

# Step 1: Raw input image
print("STEP 1: Input Image")
print("  A 224x224 color photo (JPG file)")
print("  Size: 224 x 224 x 3 (RGB) = 150,528 numbers")
print()

# Step 2: Preprocessing
print("STEP 2: Preprocessing")
steps = [
    "Resize to 224x224 pixels (standard size for most models)",
    "Normalize pixel values from 0-255 to 0-1",
    "Arrange as 3D array: [3 channels x 224 height x 224 width]",
]
for step in steps:
    print(f"  - {step}")
print()

# Step 3: CNN Layers
print("STEP 3: CNN Forward Pass")
layers = [
    ("Conv Layer 1", "224x224x3  ->  112x112x64",  "Detect edges and colors"),
    ("Conv Layer 2", "112x112x64 ->  56x56x128",   "Detect textures and patterns"),
    ("Conv Layer 3", "56x56x128  ->  28x28x256",   "Detect object parts"),
    ("Conv Layer 4", "28x28x256  ->  14x14x512",   "Detect complex features"),
    ("Global Pool",  "14x14x512  ->  512",          "Summarize all features"),
    ("Dense Layer",  "512        ->  1000",          "Combine for final prediction"),
]
for name, shape, description in layers:
    print(f"  {name:15}: {shape:25} ({description})")
print()

# Step 4: Output
print("STEP 4: Output Probabilities (top 5)")
top_5 = [
    ("golden retriever", 0.73),
    ("labrador retriever", 0.11),
    ("german shepherd", 0.05),
    ("beagle", 0.04),
    ("husky", 0.03),
]
for class_name, prob in top_5:
    bar = "=" * int(prob * 30)
    print(f"  {class_name:22}: {prob:.2f}  |{bar}|")

print()
print("Final answer: golden retriever (73% confidence)")