Project 1: Image Classification

Overview

Build convolutional neural networks from scratch for wildlife camera trap image classification. This project covers the fundamentals of image-as-data, binary classification (animal vs. no animal), and multiclass species identification.

CNN Binary Classification Multiclass Images

Notebooks

Notebook	Topic	Description
011	Image as Data	Loading and preprocessing camera trap images, understanding image tensors
013	Binary Classification	Animal vs. no animal detection using CNN
014	Multiclass Classification	Species identification across multiple categories

Key Techniques

CNN from scratch - Building convolutional layers, pooling, and fully connected layers
Custom training loops - Forward pass, loss calculation, backpropagation
Data preprocessing - Image normalization, resizing, tensor conversion
Batch processing - DataLoaders and efficient GPU utilization

Available Tools

Tool	Description
`load_proj1_info`	Get Project 1 information and available resources
`list_proj1_notebooks`	List available notebooks for image classification

Quick Start

# Get project info result = tools.get_tool('load_proj1_info')({}) # List notebooks notebooks = tools.get_tool('list_proj1_notebooks')({})