Systems and methods for product identification using image analysis and trained neural network
Disclosed are methods, systems, and non-transitory computer-readable medium for analysis of images including wearable items. For example, a method may include obtaining a first set of images, each of the first set of images depicting a product; obtaining a first set of labels associated with the first set of images; training an image segmentation neural network based on the first set of images and the first set of labels; obtaining a second set of images, each of the second set of images depicting a known product; obtaining a second set of labels associated with the second set of images; training an image classification neural network based on the second set of images and the second set of labels; receiving a query image depicting a product that is not yet identified; and performing image segmentation of the query image and identifying the product in the image by performing image analysis.
1. A computer-implemented method comprising:
receiving a query image depicting at least one wearable item;
removing at least one background image portion from the query image;
obtaining a mask image corresponding to the at least one wearable item, the obtaining based on the query image with the removed at least one background image portion;
classifying, via one or more trained neural networks, the at least one wearable item in the mask image, the classifying including determining one or more categories, one or more patterns, or one or more colors of the at least one wearable item, wherein the one or more trained neural networks was trained using training data comprising a set of labels associated with a set of images, wherein each label of the set of labels corresponds to an image of the set of images and includes information indicating a mask or classification of the at least one wearable item depicted in the corresponding image;
in response to the classifying, identifying one or more stored wearable items as one or more identified or recommended wearable items based on the mask image;
ranking the one or more identified or recommended wearable items according to a similarity score for each of the one or more identified or recommended wearable items; and
displaying the ranked one or more identified or recommended wearable items.
2. The computer-implemented method of claim 1 , further comprising:
determining the similarity score between the one or more identified or recommended wearable items and the at least one wearable item depicted in the mask image.
3. The computer-implemented method of claim 2 , further comprising:
determining one or more distance metrics between the mask image and the one or more identified or recommended wearable items.
4. The computer-implemented method of claim 3 , further comprising:
comparing the one or more distance metrics; and
assigning a higher similarity score to at least one image with a smaller distance metric.
5. The computer-implemented method of claim 1 , further comprising:
classifying the mask image according to an approximate height size or an approximate width size.
6. The computer-implemented method of claim 1 , wherein the mask image is mapped into a plurality of corresponding coordinates in an embedded space.
7. The computer-implemented method of claim 6 , wherein the embedded space is a multi-dimensional tensor space.
8. The computer-implemented method of claim 1 , further comprising:
ranking the one or more identified or recommended wearable items from a highest similarity score to a lowest similarity score in a descending order.
9. The computer-implemented method of claim 1 , further comprising:
re-ranking the one or more identified or recommended wearable items based on at least one of a pattern spatial frequency, a hue, or a saturation.
10. A computer system comprising:
a data storage device storing processor-readable instructions; and
a processor configured to execute instructions to perform a method including:
receiving a query image depicting at least one wearable item;
removing at least one background image portion from the query image;
obtaining a mask image corresponding to the at least one wearable item, the obtaining based on the query image with the removed at least one background image portion;
classifying, via one or more trained neural networks, the at least one wearable item in the mask image, the classifying including determining one or more categories, one or more patterns, or one or more colors of the at least one wearable item, wherein the one or more trained neural networks was trained using training data comprising a set of labels associated with a set of images, wherein each label of the set of labels corresponds to an image of the set of images and includes information indicating a mask or classification of the at least one wearable item depicted in the corresponding image;
in response to the classifying, identifying one or more stored wearable items as one or more identified or recommended wearable items based on the mask image;
ranking the one or more identified or recommended wearable items according to a similarity score for each of the one or more identified or recommended wearable items; and
displaying the ranked one or more identified or recommended wearable items.
11. The computer system of claim 10 , wherein the method further includes:
determining the similarity score between the one or more identified or recommended wearable items and the at least one wearable item depicted in the mask image.
12. The computer system of claim 11 , further comprising:
determining one or more distance metrics between the mask image and the one or more identified or recommended wearable items.
13. The computer system of claim 12 , further comprising:
comparing the one or more distance metrics; and
assigning a higher similarity score to at least one image with a smaller distance metric.
14. The computer system of claim 10 , further comprising:
ranking the one or more identified or recommended wearable items from a highest similarity score to a lowest similarity score in a descending order.
15. A non-transitory computer-readable medium containing instructions that, when executed by a processor, cause the processor to perform a method comprising:
receiving a query image depicting at least one wearable item;
removing at least one background image portion from the query image;
obtaining a mask image corresponding to the at least one wearable item, the obtaining based on the query image with the removed at least one background image portion;
classifying, via one or more trained neural networks, the at least one wearable item in the mask image, the classifying including determining one or more categories, one or more patterns, or one or more colors of the at least one wearable item, wherein the one or more trained neural networks was trained using training data comprising a set of labels associated with a set of images, wherein each label of the set of labels corresponds to an image of the set of images and includes information indicating a mask or classification of the at least one wearable item depicted in the corresponding image;
in response to the classifying, identifying one or more stored wearable items as one or more identified or recommended wearable items based on the mask image;
ranking the one or more identified or recommended wearable items according to a similarity score for each of the one or more identified or recommended wearable items; and
displaying the ranked one or more identified or recommended wearable items.
16. The non-transitory computer-readable medium of claim 15 , wherein the mask image is mapped into a plurality of corresponding coordinates in an embedded space.
17. The non-transitory computer-readable medium of claim 16 , wherein the embedded space is a multi-dimensional tensor space.
18. The non-transitory computer-readable medium of claim 15 , the method further comprising:
ranking the one or more identified or recommended wearable items from a highest similarity score to a lowest similarity score in a descending order.
19. The non-transitory computer-readable medium of claim 16 , the method further comprising:
re-ranking the one or more identified or recommended wearable items based on at least one of a pattern spatial frequency, a hue, or a saturation.
20. The non-transitory computer-readable medium of claim 15 , the method further comprising:
determining the similarity score between the one or more identified or recommended wearable items and the at least one wearable item depicted in the mask image.