Deep learning methods for event verification and image re-purposing detection
Systems and methods herein describe accessing an image, generating a resized image, generating an image feature vector by applying an image classification neural network to the resized image, generating analysis of the image by processing the image feature vector using a machine-learning classifier trained to analyze the image feature vector, and based on the analysis, determining an event that is attributed to the image.
1. A method comprising:
accessing, using one or more processors, an image, the image being attributed to a first specific event designated by a first time period and a first location;
generating a resized image using the image;
generating an image feature vector by applying an image classification neural network to the resized image, the image classification neural network comprising a convolutional layer that outputs to a fully-connected layer that generates the image feature vector;
generating analysis of the image by processing the image feature vector using a machine-learning classifier trained to analyze the image feature vector;
based on the analysis, determining a second specific event designated by a second time period and second location that is attributed to the image, the second specific event being different than the first specific event; and
authenticating the image based on a probability that the image was captured at the second specific event instead of the first specific event.
2. The method of claim 1 further comprising:
storing the generated image feature vector in a database.
3. The method of claim 1 , wherein the first specific event comprises a pre-defined historic event.
4. The method of claim 1 , wherein the image classification neural network comprises a pre-trained convolutional neural network trained using a set of images.
5. The method of claim 1 , wherein the image classification neural network is a fine-tuned convolutional neural network trained using a database of images, the database of images corresponding to pre-defined historic events.
6. The method of claim 1 , wherein generating the analysis of the image further comprises:
accessing an image feature vector database, the image feature vector database comprising a plurality of image feature vectors; and
identifying a matching feature vector from the plurality of image feature vectors, the matching feature vector matching to the generated image feature vector.
7. The method of claim 1 , wherein the machine-learning classifier is a nearest neighbor classifier.
8. The method of claim 1 , wherein the convolutional layer outputs a multi-dimensional dataset comprising a plurality of probabilities that the image belongs to a plurality of pre-defined image classes.
9. The method of claim 8 , further comprising:
receiving, by the fully-connected layer, the multi-dimensional dataset; and
converting the multi-dimensional dataset to a single dimensional vector.
10. A system comprising:
a processor; and
a memory storing instructions that, when executed by the processor, configure the processor to perform operations comprising:
accessing an image, the image being attributed to a first specific event designated by a first time period and a first location;
generating a resized image using the image;
generating an image feature vector by applying an image classification neural network to the resized image, the image classification neural network comprising a convolutional layer that outputs to a fully connected layer that generates the image feature vector;
generating analysis of the image by processing the image feature vector sing a machine-learning classifier trained to analyze the image feature vector;
based on the analysis, determining a second specific event designated by a second time period and second location that is attributed to the image, the second specific event being different than the first specific event; and
authenticating the image based on a probability at the image was captured at the second specific event instead of the first specific event.
11. The system of claim 10 wherein the operations further comprise:
storing the generated image feature vector in a database.
12. The system of claim 10 , wherein the second specific event comprises a pre-defined historic event.
13. The system of claim 10 , wherein the image classification neural network comprises a pre-trained convolutional neural network trained use a set of images.
14. The system of claim 10 , wherein the image classification neural network comprises a fine-tuned convolutional neural network trained using a database of images corresponding to pre-defined historic events.
15. The system of claim 10 , wherein generating the analysis of the image further comprises:
accessing an image feature vector database, the image feature vector database comprising a plurality of image feature vectors; and
identifying a matching feature vector from the plurality of image feature vectors, the matching feature vector matching to the generated image feature vector.
16. The system of claim 10 , wherein the machine-learning classifier comprises a nearest neighbor classifier.
17. The system of claim 10 , wherein the convolutional layer outputs a multi-dimensional dataset, the multi-dimensional dataset comprising a plurality of probabilities that the image belongs to a plurality of pre-defined image classes.
18. The system of claim 17 , wherein the operations further comprise:
receiving, by the fully connected layer, the multi-dimensional dataset; and
converting the multi-dimensional dataset to a single dimensional vector.
19. A non-transitory computer-readable storage medium, the non-transitory computer-readable storage medium including instructions that when executed by a computer, cause the computer to perform operations comprising:
accessing an image, the image being attributed to a first specific event designated by a first time period and a first location;
generating a resized image using the image;
generating an image feature vector by applying an image classification neural network to the resized image, the image classification neural network comprising a convolutional layer that outputs to a fully connected layer that generates the image feature vector;
generating analysis of the image by processing the image feature vector using a machine-learning classifier trained to analyze the image feature vector;
based on the analysis, determining a second specific event designated by a second time period and second location that is attributed to the image, the second specific event being different than the first specific event; and
authenticating the image based on a probability that the image was captured at the second specific event instead of the first specific event.
20. The non-transitory computer-readable storage medium of claim 19 wherein the operations further comprise:
storing the generated image feature vector in a database.