Object detection in videos

Patent metadata

Field	Value
Publication number	US-9830503-B1
Application number	US-201514984387-A
Country	US
Kind code	B1
Filing date	Dec 30, 2015
Priority date	Dec 31, 2014
Publication date	Nov 28, 2017
Grant date	Nov 28, 2017

Abstract

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for receiving a video feed of a scene. Selecting a first portion of the video feed and a second portion of the video feed based on a probability of an object being present in the first portion of the video feed compared to a probability of the object being present in the second portion of the video feed. Processing a first portion of the video feed using a first detection algorithm to detect the object in the first portion of the video feed. Processing a second portion of the video feed using a second detection algorithm to detect the object in the second portion of the video feet, where the first detection algorithm is different from the second detection algorithm.

First claim

What is claimed is:

1. A computer-implemented method for selectively applying multiple object detection algorithms, the method being executed by one or more processors and comprising:
receiving, by the one or more processors, a video feed of a scene;
selecting, by the one or more processors, a first portion of the video feed and a second portion of the video feed based on a probability of an object being present in the first portion of the video feed compared to a probability of the object being present in the second portion of the video feed, wherein the first portion of the video feed is a first region of the scene in a series of frames and the second portion of the video feed is a second region of the scene in the series of frames;
determining, by the one or more processors, a budget of available computing resources;
selecting, by the one or more processors, a size of the first region and a size of the second region based on the budget of available computing resources;
processing, by the one or more processors, a first portion of the video feed using a first detection algorithm to detect the object in the first portion of the video feed; and
processing, by the one or more processors, a second portion of the video feed using a second detection algorithm to detect the object in the second portion of the video feed,
wherein the first detection algorithm is different from the second detection algorithm.

2. The method of claim 1 , wherein the second detection algorithm is more computationally intensive than the first detection algorithm and the probability of the object being present in the second portion of the video feed is greater than the probability of the object being present in the first portion of the video feed.

3. The method of claim 1 , wherein the first portion of the video feed includes a first set of frames and the second portion of the video feed includes a second set of frames.

4. The method of claim 2 , further comprising:
obtaining a probability model of the scene in the video feed, wherein the probability model is based on historical detections of the object within the scene; and
determining the probability of the object being present in the first portion of the video feed and the probability of the object being present in the second portion of the video feed based on the probability model of the scene in the video feed.

5. The method of claim 4 , wherein the probability model includes probabilities of detecting the object mapped to regions within the scene.

6. The method of claim 4 , further comprising:
detecting the object in the video feed; and
in response to detecting the object, modifying the probability model based on a region within the scene in which object was detected.

7. The method of claim 4 , wherein the probability model includes a spatial probability model, that maps probabilities of detecting the object to regions within the scene, and a temporal probability model that maps probabilities of detecting the object to time periods.

8. The method of claim 2 , further comprising:
detecting motion in a first frame of the video feed; and
determining, based on the detected motion in the first frame, the probability of the object being present in a second frame of the video, and
wherein the second frame is selected as the second portion of the video feed.

9. The method of claim 1 , further comprising detecting the object in a region of a first frame of the video feed, and
wherein selecting the first portion of the video feed and the second portion of the video feed comprises selecting the second portion of the video feed to include a region of a second frame that is proximate to the region of the first frame, wherein the second frame is subsequent in time to the first frame.

10. The method of claim 1 , wherein the first portion includes the second portion.

11. The method of claim 1 , wherein the object is a face, wherein the first detection algorithm is a first facial detection algorithm, and wherein the second detection algorithm is a second facial detection algorithm.

12. The method of claim 1 , wherein the video feed is a live video feed.

13. The method of claim 1 , wherein the video feed is a recorded video feed.

14. A system for selectively applying multiple object detection algorithms, the system comprising:
one or more computers; and
a computer-readable medium coupled to the one or more computers having instructions stored thereon which, when executed by the one or more computers, cause the one or more computers to perform operations comprising:
receiving a video feed of a scene;
selecting a first portion of the video feed and a second portion of the video feed based on a probability of an object being present in the first portion of the video feed compared to a probability of the object being present in the second portion of the video feed, wherein the first portion of the video feed includes a first region of the scene in a series of frames and the second portion of the video feed includes a second region of the scene in the series of frames;
determining a budget of available computing resources;
selecting a size of the first region and a size of the second region based on the budget of available computing resources;
processing a first portion of the video feed using a first detection algorithm to detect the object in the first portion of the video feed; and
processing a second portion of the video feed using a second detection algorithm to detect the object in the second portion of the video feed,
wherein the first detection algorithm is different from the second detection algorithm.

15. The system of claim 14 , wherein the second detection algorithm is more computationally intensive than the first detection algorithm and the probability of the object being present in the second portion of the video feed is greater than the probability of the object being present in the first portion of the video feed.

16. The system of claim 15 , wherein the operations further comprise:
obtaining a probability model of the scene in the video feed, wherein the probability model is based on historical detections of the object within the scene; and
determining the probability of the object being present in the first portion of the video feed and the probability of the object being present in the second portion of the video feed based on the probability model of the scene in the video feed.

17. A non-transient computer readable storage device storing instructions that, when executed by at least one processor, cause the at least one processor to perform operations comprising:
receiving a video feed of a scene;
selecting a first portion of the video feed and a second portion of the video feed based on a probability of an object being present in the first portion of the video feed compared to a probability of the object being present in the second portion of the video feed, wherein the first portion of the video feed includes a first region of the scene in a series of frames and the second portion of the video feed includes a second region of the scene in the series of frames;
determining a budget of available computing resources;
selecting a size of the first region and a size of the second region based on the budget of available computing resources;
processing a first portion of the video feed using a first detection algorithm to detect the object in the first portion of the video feed; and
processing a second portion of the video feed using a second detection algorithm to detect the object in the second portion of the video

Assignees

• Morphotrust Usa Llc

Inventors

• Martin Brian
• Bertan Brian

Classifications

• G06V10/764

  using classification, e.g. of video objects · CPC title

• G06V40/166

  using acquisition arrangements · CPC title

• G06V20/52 Primary

  Surveillance or monitoring of activities, e.g. for recognising suspicious objects (recognising microscopic objects G06V20/69) · CPC title

• G06T2207/30201

  Face · CPC title

• G06T7/20

  Analysis of motion (motion estimation for coding, decoding, compressing or decompressing digital video signals H04N19/43, H04N19/51) · CPC title