SCL DOT

An iOS field-dashboard app for the Trapjaw real-time insect rotoscoping pipeline.

What it does

SCL DOT runs the Trapjaw computer-vision engine directly on an iPhone, capturing live video and detecting, tracking, and cropping flying insects in real time using Apple Metal compute shaders. The app presents a minimal heads-up display of pipeline metrics so you can monitor performance in the field.

Features

• Live metrics dashboard — camera FPS, processing FPS, average pipeline latency, GPU time, active track count, and pending uploads
• Background capture — manually trigger a background image capture for the segmentation model
• Video clip recording — record and upload short video clips on demand
• Operating hours — automatic start / pause based on a configurable time window
• Cool-down & status indicators — visual feedback for warmup, cooling, paused, and offline states
• Settings sheet — in-app configuration

Architecture

• SwiftUI front end (TrapjawApp, ContentView, SettingsView)
• TrapjawProcessor (Processing/TrapjawProcessor.swift) — bridges the SwiftUI layer to the underlying Trapjaw C++ pipeline
• Trapjaw submodule (trapjaw/) — the core detection/tracking library
• Pre-built libraries (lib/) — libtrapjaw.a for iOS device and simulator, plus trapjaw.metallib

Requirements

• iOS 18.0+
• Xcode 16+
• CMake 3.20+ (for rebuilding the Trapjaw library)
• An iPhone with a Metal-capable GPU

Building

1. Clone the repo and initialize submodules:

   git clone --recurse-submodules https://github.com/MIT-Senseable-City-Lab/dot.git

2. Build the Trapjaw library and Metal shaders: ./build-trapjaw.sh all
3. Open TrapjawApp.xcodeproj in Xcode and build for your target device. License MIT