Computer Vision for Beginners: Building an Image Classifier

Computer vision enables AI to understand and process visual information. This guide walks through building your first image classifier.

What is Computer Vision?

Computer vision is the field of AI that trains computers to interpret and understand the visual world. Applications include:

• Image classification (what's in this photo?)
• Object detection (where are the objects?)
• Image segmentation (which pixels belong to which object?)
• Face recognition (who is this person?)

Setting Up

pip install tensorflow matplotlib numpy pillow

Loading Data

We'll use the CIFAR-10 dataset (60,000 32x32 color images in 10 classes):

import tensorflow as tf
from tensorflow import keras

(x_train, y_train), (x_test, y_test) = keras.datasets.cifar10.load_data()

# Normalize pixel values
x_train = x_train.astype('float32') / 255
x_test = x_test.astype('float32') / 255

Building a Convolutional Neural Network

model = keras.Sequential([
    keras.layers.Conv2D(32, (3,3), activation='relu', input_shape=(32,32,3)),
    keras.layers.MaxPooling2D(2,2),
    keras.layers.Conv2D(64, (3,3), activation='relu'),
    keras.layers.MaxPooling2D(2,2),
    keras.layers.Flatten(),
    keras.layers.Dense(128, activation='relu'),
    keras.layers.Dense(10, activation='softmax')
])

model.compile(
    optimizer='adam',
    loss='sparse_categorical_crossentropy',
    metrics=['accuracy']
)

Training

history = model.fit(
    x_train, y_train,
    batch_size=64,
    epochs=10,
    validation_data=(x_test, y_test)
)

Using Transfer Learning

For better results with less data, use a pre-trained model:

base_model = keras.applications.MobileNetV2(
    input_shape=(224,224,3),
    include_top=False,
    weights='imagenet'
)
base_model.trainable = False  # Freeze base layers

model = keras.Sequential([
    keras.layers.Resizing(224,224),
    base_model,
    keras.layers.GlobalAveragePooling2D(),
    keras.layers.Dense(128, activation='relu'),
    keras.layers.Dense(10, activation='softmax')
])

Object Detection with YOLO

For detecting multiple objects in images, use YOLO (You Only Look Once):

pip install ultralytics
yolo predict model=yolov8n.pt source='image.jpg'

Real-World Application: Moroccan Agriculture

A team of Moroccan researchers built a vision system using transfer learning that:

• Identifies crop diseases from smartphone photos
• Achieves 94% accuracy on local crop varieties
• Works offline on mobile devices
• Helps farmers diagnose problems early

Next Steps

• Try real-time video processing
• Build a mobile app with on-device vision
• Explore image generation with diffusion models

Our AI Builder Bootcamp covers computer vision with real-world projects.