Identifying an object within content

Abstract

A method for identifying an object within a video sequence, wherein the video sequence comprises a sequence of images, wherein the method comprises, for each of one or more images of the sequence of images: using a first neural network to determine whether or not an object of a predetermined type is depicted within the image; and in response to the first neural network determining that an object of the predetermined type is depicted within the image, using an ensemble of second neural networks to identify the object determined as being depicted within the image.

1 . A method for identifying an object within a video sequence, wherein the video sequence comprises a sequence of images, wherein the method comprises, for each of one or more images of the sequence of images:
 using a first neural network to determine whether or not an object of a predetermined type is depicted within the image; and   in response to the first neural network determining that an object of the predetermined type is depicted within the image, using an ensemble of second neural networks to identify the object determined as being depicted within the image.   
     
     
         2 . The method of  claim 1 , wherein the first neural network and/or one or more of the second neural networks is a convolutional neural network or a deep convolutional neural network. 
     
     
         3 . (canceled) 
     
     
         4 . The method of  claim 1 , wherein using a first neural network to determine whether or not an object of a predetermined type is depicted within the image comprises:
 generating a plurality of candidate images from the image;   using the first neural network to determine, for each of the candidate images, an indication of whether or not an object of the predetermined type is depicted in said candidate image; and   using the indications to determine whether or not an object of the predetermined type is depicted within the image.   
     
     
         5 . The method of  claim 4 , wherein one or more of the candidate images is generated from the image by performing one or more geometric transformations on an area of the image. 
     
     
         6 . The method of  claim 1 , wherein the predetermined type is a logo. 
     
     
         7 . The method of  claim 1 , wherein the predetermined type is a face or a person. 
     
     
         8 . The method of  claim 1 , comprising associating metadata with the image based on the identified object. 
     
     
         9 . The method of  claim 6 , comprising:
 obtaining the video sequence from a source; and   determining unauthorized use of the video sequence based on identifying that the logo is depicted within one or more images of the video sequence.   
     
     
         10 . The method of  claim 9 , wherein the logo is one of a plurality of predetermined logos. 
     
     
         11 . A method for identifying an object within an amount of content, the method comprising:
 using a first neural network to determine whether or not an object of a predetermined type is depicted within the amount of content; and   in response to the first neural network determining that an object of the predetermined type is depicted within the amount of content, using an ensemble of second neural networks to identify the object determined as being depicted within the amount of content.   
     
     
         12 . The method of  claim 11 , wherein the amount of content is one of: (a) an image; (b) an image of a video sequence that comprises a sequence of images; and (c) an audio snippet. 
     
     
         13 . The method of  claim 11 , wherein the first neural network and/or one or more of the second neural networks is a convolutional neural network or a deep convolutional neural network. 
     
     
         14 . (canceled) 
     
     
         15 . The method of  claim 11 , wherein using a first neural network to determine whether or not an object of a predetermined type is depicted within the amount of content comprises:
 generating a plurality of content candidates from the amount of content;   using the first neural network to determine, for each of the content candidates, an indication of whether or not an object of the predetermined type is depicted in said content candidate; and   using the indications to determine whether or not an object of the predetermined type is depicted within the amount of content.   
     
     
         16 . The method of  claim 15 , wherein one or more of the content candidates is generated from the amount of content by performing one or more geometric transformations on a portion of the amount of content. 
     
     
         17 . The method of  claim 11 , wherein the amount of content is an audio snippet and the predetermined type is one of: a voice; a word; a phrase. 
     
     
         18 . The method of  claim 11 , comprising associating metadata with the amount of content based on the identified object. 
     
     
         19 . An apparatus comprising one or more processors, the one or more processors being arranged to carry out identification of an object within an amount of content, said identification comprising:
 using a first neural network to determine whether or not an object of a predetermined type is depicted within the amount of content; and   in response to the first neural network determining that an object of the predetermined type is depicted within the amount of content, using an ensemble of second neural networks to identify the object determined as being depicted within the amount of content.   
     
     
         20 . (canceled) 
     
     
         21 . A non-transitory computer-readable medium storing a computer program which, when executed by one or more processors, causes the one or more processors to carry out identification of an object within an amount of content, said identification comprising:
 using a first neural network to determine whether or not an object of a predetermined type is depicted within the amount of content; and   in response to the first neural network determining that an object of the predetermined type is depicted within the amount of content, using an ensemble of second neural networks to identify the object determined as being depicted within the amount of content.   
     
     
         22 . The apparatus of  claim 19 , wherein the amount of content is one of: (a) an image (b) an image of a video sequence that comprises a sequence of images; and (c) or an audio snippet. 
     
     
         23 . The non-transitory computer-readable medium of  claim 21 , wherein the amount of content is one of: (a) an image (b) an image of a video sequence that comprises a sequence of images; and (c) or an audio snippet.