Sample AI Projects Gallery

Project Overview

The Smart Recycling Sorter helps users properly categorize waste items into Plastic, Paper, Metal, and Glass. This AI addresses the real-world problem of recycling confusion, where people are often unsure which recycling bin to use. The AI analyzes images of common household items and provides immediate classification guidance.

Training Data Strategy

• Plastic Category (85 samples): Water bottles, food containers, plastic bags, shopping bags, plastic utensils, detergent bottles - varied colors, sizes, lighting conditions
• Paper Category (82 samples): Newspapers, cardboard boxes, paper bags, magazines, envelopes, notebook paper - different textures, colors, crumpled and flat
• Metal Category (78 samples): Aluminum cans, tin cans, metal lids, foil, aerosol cans - various sizes, clean and dirty, different angles
• Glass Category (80 samples): Glass bottles, jars, drinking glasses - clear and colored glass, different sizes and shapes, various lighting

Implementation Steps

Key Features

• Real-time Classification: Instant feedback as items are shown to the camera
• Confidence Scores: Shows percentage confidence for each category
• Visual Feedback: Color-coded results (green for high confidence, yellow for medium, red for low)
• Instructions Display: Shows proper disposal instructions for each category

Project Overview

The Facial Emotion Detector recognizes four primary emotions: Happy, Sad, Surprised, and Neutral. This AI can be used in educational settings to help students understand emotional recognition, in games for emotion-based controls, or as an accessibility tool for individuals who struggle with reading facial expressions.

Training Data Strategy

• Happy (125 samples): Big smiles, small smiles, laughing, grinning - different people, angles, with/without glasses
• Sad (120 samples): Frowning, downturned mouth, sad eyes, looking down - various intensities and expressions
• Surprised (118 samples): Wide eyes, open mouth, raised eyebrows - genuine and exaggerated surprise
• Neutral (122 samples): Relaxed face, no expression, calm demeanor - looking at camera and away

Implementation Steps

Key Features

• Live Emotion Tracking: Real-time analysis of facial expressions
• Emoji Overlay: Displays corresponding emoji for detected emotion
• Confidence Meter: Visual bar graph showing confidence levels
• Data Logging: Records emotion changes over time
• Sound Effects: Audio feedback matching detected emotions

Project Overview

This AI identifies sounds from five different musical instruments: Guitar, Piano, Drums, Violin, and Human Voice. Perfect for music education, this tool helps students learn to distinguish between different instrument timbres and could be used in interactive music games or as an aid for young musicians.

Training Data Strategy

• Guitar (78 samples): Acoustic strumming, electric riffs, fingerpicking, chords, single notes - different playing styles
• Piano (82 samples): Single notes, chords, arpeggios, low and high registers - various dynamics and tempos
• Drums (76 samples): Bass drum, snare, cymbals, tom-toms, full beats - different patterns and intensities
• Violin (75 samples): Long notes, short notes, pizzicato, different bow pressures - high and low strings
• Voice (80 samples): Singing different pitches, humming, vowel sounds - male and female voices, different ranges

Implementation Steps

Key Features

• Real-time Audio Analysis: Identifies instruments as they play
• Visual Waveform Display: Shows audio input with frequency spectrum
• Instrument Icons: Displays corresponding instrument image
• Background Music Support: Can identify instruments in mixed audio
• Educational Mode: Shows fun facts about each identified instrument

Project Overview

This innovative AI controller recognizes five hand gestures to control a game: Thumbs Up (jump), Peace Sign (left), Fist (right), Open Palm (stop), and OK Sign (select). Integrated with Scratch, this creates a touchless gaming experience that demonstrates how AI can transform human movement into digital commands.

Training Data Strategy

• Thumbs Up (95 samples): Various hand angles, distances from camera, left and right hands
• Peace Sign (92 samples): Different orientations, hand positions, finger spreads
• Fist (88 samples): Tight and loose fists, different angles, hand rotations
• Open Palm (94 samples): Fingers together and spread, facing camera and angled
• OK Sign (91 samples): Different finger circle sizes, various hand orientations

Implementation Steps

Scratch Game Integration

The Teachable Machine model was integrated into a Scratch platform game where:

• Thumbs Up: Makes the character jump
• Peace Sign (left hand): Moves character left
• Fist (right hand): Moves character right
• Open Palm: Character stops and crouches
• OK Sign: Activates special ability

Project Overview

This mobile app helps gardeners identify plant health issues by analyzing leaf photos. The AI classifies plants as Healthy, Drought-Stressed, Disease-Infected, or Pest-Damaged, then provides specific care recommendations. This demonstrates how AI can be used in agriculture and home gardening to diagnose problems early.

Training Data Strategy

• Healthy (105 samples): Vibrant green leaves, no discoloration, various plant species, different leaf shapes
• Drought-Stressed (98 samples): Wilted leaves, brown edges, drooping, curled leaves, various severities
• Disease-Infected (102 samples): Spots, mold, discoloration patterns, fungal growth, bacterial infections
• Pest-Damaged (95 samples): Holes in leaves, chewed edges, visible insects, pest damage patterns

Implementation Steps

Key App Features

• Camera Integration: Take photo directly from app or select from gallery
• Instant Analysis: AI processes image and provides diagnosis within seconds
• Care Recommendations: Displays specific action items based on diagnosis
• Confidence Display: Shows how certain the AI is about its classification
• History Log: Saves previous diagnoses for tracking plant health over time
• Share Function: Allows users to share results with other gardeners