How kingtmata/btl-ai-camera-cd8 works
Classroom AI is a local classroom-image and video detection/tracking application with separate video, terminal, and Streamlit interfaces, plus a dataset and model-training workflow. The main video path accepts a user-selected video, runs detection (optionally tracking), and produces annotated playback or an exported MP4. A separate research branch evaluates inference variants and validation campaigns. Edges are limited to README-documented behavior and relationships visible in the supplied source excerpts or SOURCE INDEX; unsampled details are not asserted.
Main entry points: pipeline.py Video pipeline · app.py Web interface · worker.py Training worker · detector.py Object detector
Components 28
User interfaces
- Video entry point
src/classroom_ai/video/__main__.py - Terminal UI
src/classroom_ai/interfaces/terminal.py - Video controls
src/classroom_ai/interfaces/terminal_process.py - Web interface
src/classroom_ai/interfaces/web/app.py - Image preview
src/classroom_ai/interfaces/web/preview.py - Training interface
src/classroom_ai/interfaces/web/training.py
Video processing
- Video pipeline
src/classroom_ai/video/pipeline.py - Media I/O
src/classroom_ai/video/media.py - Video core
src/classroom_ai/video/core.py - Person tracking
src/classroom_ai/video/tracked_processing.py - Video rendering
src/classroom_ai/video/rendering.py - Video playback
src/classroom_ai/video/playback.py - Run metrics
src/classroom_ai/video/metrics.py
Detection and inference
- Object detector
src/classroom_ai/inference/detector.py - Webcam inference
src/classroom_ai/inference/camera.py - Device selection
src/classroom_ai/inference/devices.py - Inference variants
src/classroom_ai/inference/variants.py
Datasets and training
- Dataset validation
src/classroom_ai/datasets.py - Class definitions
src/classroom_ai/classes.py - Training runs
src/classroom_ai/training/manager.py - Training worker
src/classroom_ai/training/worker.py - Model evaluation
src/classroom_ai/evaluation.py - Kaggle export
src/classroom_ai/training/kaggle_export.py
Research evaluation
- Experiment campaigns
src/classroom_ai/experiments/experiment_campaign.py - Validation replay
src/classroom_ai/experiments/validation_replay.py - Improvement evaluation
src/classroom_ai/experiments/e3_improvement_eval.py
Other components
- Local user
- Video input
Connections 35
- Local user → Video entry point: starts
README.md - Local user → Video input: selects
README.md - Video input → Video pipeline: processes
README.md - Video entry point → Terminal UI: offers UI
src/classroom_ai/video/__main__.py - Video entry point → Video controls: dispatches
src/classroom_ai/video/__main__.py - Video entry point → Video pipeline: runs
src/classroom_ai/video/__main__.py - Terminal UI → Video controls: uses controls
src/classroom_ai/interfaces/terminal.py - Video pipeline → Media I/O: reads and exports
src/classroom_ai/video/pipeline.py - Video pipeline → Video core: uses
src/classroom_ai/video/pipeline.py - Video pipeline → Object detector: runs detection
src/classroom_ai/video/pipeline.py - Video pipeline → Person tracking: dispatches tracking
src/classroom_ai/video/pipeline.py - Video pipeline → Run metrics: records metrics
src/classroom_ai/video/pipeline.py - Video pipeline → Video playback: supports playback
src/classroom_ai/video/pipeline.py - Person tracking → Object detector: uses
src/classroom_ai/video/tracked_processing.py - Person tracking → Media I/O: uses
src/classroom_ai/video/tracked_processing.py - Person tracking → Video rendering: uses
src/classroom_ai/video/tracked_processing.py - Local user → Web interface: opens
README.md - Web interface → Webcam inference: controls webcam
src/classroom_ai/interfaces/web/app.py - Web interface → Object detector: uses
src/classroom_ai/interfaces/web/app.py - Web interface → Image preview: renders preview
src/classroom_ai/interfaces/web/app.py - Web interface → Training interface: opens training
src/classroom_ai/interfaces/web/app.py - Web interface → Training runs: checks active run
src/classroom_ai/interfaces/web/app.py - Image preview → Object detector: draws detections
src/classroom_ai/interfaces/web/preview.py - Training interface → Training runs: manages runs
src/classroom_ai/interfaces/web/training.py - Training interface → Kaggle export: exports training data
src/classroom_ai/interfaces/web/training.py - Training runs → Training worker: spawns worker
src/classroom_ai/training/worker.py - Training worker → Training runs: validates run
src/classroom_ai/training/worker.py - Training worker → Dataset validation: validates release
src/classroom_ai/training/worker.py - Training worker → Model evaluation: uses protocol
src/classroom_ai/training/worker.py - Training worker → Class definitions: checks classes
src/classroom_ai/training/worker.py - Model evaluation → Dataset validation: uses data
src/classroom_ai/evaluation.py - Model evaluation → Object detector: uses detector
src/classroom_ai/evaluation.py - Experiment campaigns → Inference variants: uses variants
- Validation replay → Inference variants: replays variants
src/classroom_ai/experiments/validation_replay.py - Improvement evaluation → Experiment campaigns: evaluates campaign