DeepCount

DeepCount uses AI to detect people in your camera view or recorded videos and counts how many cross the lines you draw on screen, separately for IN and OUT. Run foot-traffic surveys and measure the flow of people in front of a store with nothing but an iPhone or iPad. FEATURES • Set count lines just by tracing with your finger (up to 4 lines) • Automatic IN / OUT counts for each line • Four video sources - Live: count on the spot with the device camera - Recorded (Photos): analyze a video from your photo library - Recorded (Files): analyze a video from the Files app - Stream: analyze video from a network camera and more (HLS / MP4 / MJPEG) • Tiled inference detects people who appear small in the distance • Recorded videos are analyzed frame by frame and logged with the recording time • Export counting data as CSV files in a ZIP (AirDrop, Save to Files, Mail, and more) • Import your own AI models (.mlmodel / .mlpackage / .zip) • Available in English and Japanese GREAT FOR • Counting visitors and foot traffic at stores and facilities • Understanding crowd flow at events and in passages • Tallying videos you have already recorded EXPORTED DATA • data_*.csv: detection records (split every 6 hours) • lineCount_*.csv: line crossing records (line number, IN/OUT, time) PRIVACY All video analysis happens on your device. Video and counting data are never sent to an external server. Location is used to record where the counting took place (latitude and longitude). TIPS • Mount the device on a tripod or similar so the video does not shake. • Draw lines across the path people walk, covering the full width of the passage. For details, tap "How to Use" on the top screen of the app.