{
  "patent_number": "US 12738117",
  "country": "US",
  "title": "Estimating Queue Lengths and Wait Times Using Camera Images",
  "original_title": "Methods of estimating a throughput of a resource, a length of a queue associated with the resource and/or a wait time of the queue",
  "summary": "This patent describes a system that uses cameras to count people in a line and estimate how long they will have to wait, then sends that information to their phones.",
  "what_it_does": "This patent outlines a method for automatically estimating the number of people in a queue and their expected wait time at a venue. It works by receiving multiple images, each showing a different part, or 'segment,' of the queue. A central server then processes these images using two different techniques: for some images, it estimates the number of people by analyzing how many pixels correspond to guests; for others, it directly detects and counts individual people visible in the image. Once the total number of guests in the queue is estimated, the system calculates the current wait time or the earliest entry time, and sends this information as a notification to the devices of venue guests. For example, a theme park guest could receive a notification on their phone saying, 'The wait for the roller coaster is 30 minutes,' based on real-time camera analysis.",
  "what_it_does_not_cover": [
    "Does not cover systems that rely solely on manual counting or non-image-based sensors to estimate queue length.",
    "Does not cover systems that only use a single, wide-angle image to estimate the entire queue without segmenting it.",
    "Does not cover systems that exclusively use pixel estimation without also detecting and counting individual guests in at least some images.",
    "Does not cover systems that estimate queue metrics but do not send notifications to individual guest devices.",
    "Does not cover methods of estimating wait times that do not involve image analysis of the queue itself."
  ],
  "filed": "2023-07-12",
  "granted": "2026-09-15",
  "expires": null,
  "status": "active",
  "holder": null,
  "holder_url": null,
  "inventors": [],
  "times_cited": 0,
  "tags": [
    "consumer_electronics",
    "software",
    "telecommunications",
    "retail",
    "entertainment",
    "ai_ml"
  ],
  "abstract": "Aspects of the disclosure relate to a method of estimating a length of a queue associated with a resource within a venue. Images of the queue may be received, each image corresponding to a segment of the queue. A server may estimate a number of venue guests in the queue by: for each of a first subset of the images, estimating a number of pixels of the image that correspond to venue guests in the queue; and for each of a second subset of the images, detecting and counting individual venue guests visible in the image. A notification including the estimated number of venue guests in the queue, an estimated current wait time, and/or a time at which a venue guest joining the queue is permitted entry to the resource, may be sent to devices corresponding to venue guests.",
  "url": "https://patentbrief.org/patent/us/12738117/methods-of-estimating-a-throughput-of-a-resource-a-length-of-a-queue",
  "markdown_url": "https://patentbrief.org/patent/us/12738117/methods-of-estimating-a-throughput-of-a-resource-a-length-of-a-queue/md",
  "google_patents_url": "https://patents.google.com/patent/US12738117",
  "relatedPatents": []
}