System and method for automated electronic catalogue management and electronic image quality assessment

Abstract

In various examples, a system receives image data characterizing an image of an item. Additionally, the system implements a first set of operations and a second set of operations. In some examples, the first set of operations includes performing a structural similarity analysis of the item, based on the image data, and determining a structural similarity score based on the structural similarity analysis of the item. In other examples, the second set of operations includes generating a plurality of derivative images by applying a plurality of distortions to the image of the item, extracting one or more features based at least on the plurality of derivative images, and determining the quality of the image based at least on the extracted one or more features and the structural similarity score.

What is claimed is: 
     
         1 . A system comprising:
 one or more processors; and   a memory resource storing a set of instructions, that when executed by the one or more processors, causes the one or more processors to:
 receive image data characterizing an image of an item; 
 determine a quality of the image based on one or more features extracted from the image based on a plurality of derivative images and a structural similarity score, wherein the plurality of derivative images are generated by applying at least one distortion to the image of the item; and 
 determine an orientation of the image based on a detected classifier loss. 
   
     
     
         2 . The system of  claim 1 , wherein the quality of the image is determined by applying a regression model to the one or more features and the structural similarity score. 
     
     
         3 . The system of  claim 2 , wherein the regression model is a ridge regression model. 
     
     
         4 . The system of  claim 1 , wherein the one or more features are extracted by a convolutional neural network. 
     
     
         5 . The system of  claim 4 , wherein the detected classifier loss is determined by a classifier loss function configured to receive an output of the convolutional neural network. 
     
     
         6 . The system of  claim 5 , wherein the output of the convolutional neural network includes kernel principal component analysis features and image embeddings. 
     
     
         7 . The system of  claim 1 , wherein the at least one distortion is one of a mean blur, a Gaussian blur, or a bilateral blur. 
     
     
         8 . The system of  claim 1 , wherein the structural similarity score is determined based on changes luminance, contrast, and structure of the image. 
     
     
         9 . A computer-implemented method, comprising:
 receiving image data characterizing an image of an item;   determining a quality of the image based on one or more features extracted from the image based on a plurality of derivative images and a structural similarity score, wherein the plurality of derivative images are generated by applying at least one distortion to the image of the item; and   determining an orientation of the image based on a detected classifier loss.   
     
     
         10 . The computer-implemented method of  claim 9 , wherein the quality of the image is determined by applying a regression model to the one or more features and the structural similarity score. 
     
     
         11 . The computer-implemented method of  claim 10 , wherein the regression model is a ridge regression model. 
     
     
         12 . The computer-implemented method of  claim 9 , wherein the one or more features are extracted by a convolutional neural network. 
     
     
         13 . The computer-implemented method of  claim 12 , wherein the detected classifier loss is determined by a classifier loss function configured to receive an output of the convolutional neural network. 
     
     
         14 . The computer-implemented method of  claim 13 , wherein the output of the convolutional neural network includes kernel principal component analysis features and image embeddings. 
     
     
         15 . The computer-implemented method of  claim 9 , wherein the at least one distortion is one of a mean blur, a Gaussian blur, or a bilateral blur. 
     
     
         16 . The computer-implemented method of  claim 9 , wherein the structural similarity score is determined based on changes luminance, contrast, and structure of the image. 
     
     
         17 . A non-transitory computer-readable medium storing instructions, that when executed by one or more processors, causes the one or more processors to:
 receive image data characterizing an image of an item;   extract one or more features from the image, wherein the one or more features are extracted by a convolutional neural network;   determine a quality of the image based on the one or more features extracted from the image based on a plurality of derivative images and a structural similarity score, wherein the plurality of derivative images are generated by applying at least one distortion to the image of the item; and   determine an orientation of the image based on a detected classifier loss based on an output of the convolutional neural network.   
     
     
         18 . The non-transitory computer-readable medium of  claim 17 , wherein the output of the convolutional neural network includes kernel principal component analysis features and image embeddings. 
     
     
         19 . The non-transitory computer-readable medium of  claim 17 , wherein the quality of the image is determined by applying a regression model to the one or more features and the structural similarity score. 
     
     
         20 . The non-transitory computer-readable medium of  claim 17 , wherein the at least one distortion is one of a mean blur, a Gaussian blur, or a bilateral blur.