Method and a system for recognition of data in one or more images
The present disclosure relates to a method and system for recognition of data in one or more images. The method receives and segments the one or more images to identify segmented objects. Further, the method generates an object relationship data for each of the segmented objects and determines a knowledge base representation of the object relationship data based on defined features. Furthermore, a Recurrent Neural Network (RNN) is trained based on the knowledge base representation to determine an appropriate Neural Network (NN) having optimum confidence score. Based on the appropriate NN selected, the objects in the input image is predicted and transmitted to external systems for decision making. Thus, enabling accurate text/object identification for the images having different background variations using the dynamic selection of NN, thereby facilitating more effective decision making.
1. A method for recognition of data in one or more images, the method comprising:
receiving, by a processor of a data recognition system, the one or more images from an image sensor coupled with the processor;
segmenting, by the processor, the one or more images to identify one or more segmented objects in the one or more images;
generating, by the processor, an object relationship data for each of the segmented objects, each of the segmented objects comprising at least one or more defined features and a confidence score;
determining, by the processor, a knowledge base representation of the object relationship data based on the one or more defined features of the one or more segmented objects, wherein the knowledge base representation comprises one or more defined feature nodes, one or more indent nodes and a plurality of links defining a relationship between the one or more defined feature nodes and at least one indent node;
training, by the processor, a Recurrent Neural Network (RNN) based on the knowledge base representation to generate a trained RNN; and
determining, by the processor, an appropriate Neural Network (NN) based on selection of the at least one indent node having optimum confidence score and the trained RNN for data prediction and recognition, wherein determining the appropriate NN comprises selecting the NN by the RNN based on at least one defined feature node linked with the indent node representing the selected NN.
2. The method as claimed in claim 1 , wherein segmenting the one or more images comprising, the steps of:
identifying one or more trained objects and text data in the one or more images;
generating, by the processor, at least one boundary for the one or more trained objects text data identified in the one or more images; and
cropping, by the processor, the identified trained objects and text data along the at least one boundary to determine the one or more segmented objects in the one or more images.
3. The method as claimed in claim 1 , wherein generating the object relationship data comprising steps of:
training a plurality of NNs with the one or more segmented objects;
identifying at least one defined feature for each of the one or more segmented objects; and
determining the confidence score for the one or more defined features of one or more segmented objects associated with the plurality of NNs.
4. The method as claimed in claim 1 , wherein the step of determining the knowledge base representation of the object relationship data comprising steps of:
identifying the one or more indent nodes for each segmented object in the object relationship data, wherein an indent node represents a NN;
determining at least one indent node with optimum confidence score associated with each segmented object; and
mapping the at least one indent node with the one or more defined features to generate a relationship between the one or more defined feature nodes and at least one indent node, wherein a defined feature node represents a defined feature in the knowledge base representation.
5. The method as claimed in claim 1 , further comprising:
automatically updating the knowledge base representation of the object relationship data with new NN based on the one or more defined features; and
transmitting the data predicted and recognized from the selected NN to a decision-making module for further processing.
6. A system for recognition of data in one or more images, the system comprising:
at least one image sensor for capturing the one or more images;
a processor; and
a memory, communicatively coupled with the processor, wherein the memory stores processor-executable instructions, which on execution cause the processor to:
receive the one or more images from the image sensor;
segment the one or more images to identify one or more segmented objects;
generate an object relationship data for each of the segmented objects, each segmented object comprises one or more defined features and a confidence score;
determine a knowledge base representation of the object relationship data based on the one or more defined features of the one or more segmented objects, wherein the knowledge base representation comprises the one or more defined feature nodes, one or more indent nodes and a plurality of links defining the relationship between the one or more defined feature nodes and at least one indent node;
train a Recurrent Neural Network (RNN) based on the knowledge base representation to generate a trained RNN; and
determine an appropriate Neural Network (NN) based on selection of the at least one indent node having optimum confidence score and the trained RNN for data prediction and recognition, wherein the processor is configured to determine the appropriate NN comprises selecting the NN by the RNN based on at least one defined feature node linked with the indent node representing the selected NN.
7. The system as claimed in claim 6 , wherein the processor is configured to segment the one or more images by steps comprising:
identifying one or more trained objects and text data in the one or more images;
generating at least one boundary for the one or more trained objects and text data identified in the one or more images; and
cropping the identified trained objects and text data along the at least one boundary to determine the one or more segmented objects in the one or more images.
8. The system as claimed in claim 6 , wherein the processor is configured to generate the object relationship data by performing steps of:
training a plurality of NNs with the one or more segmented objects;
identifying at least one defined feature for each of the one or more segmented objects; and
determining the confidence score for the one or more defined features of one or more segmented objects associated with the plurality of NNs.
9. The system as claimed in claim 6 , wherein the processor is configured to determine the knowledge base representation of the object relationship data by performing steps of;
identifying the one or more indent nodes for each segmented object in the object relationship data, wherein an indent node represents a NN;
determining at least one indent node with optimum confidence score associated with each segmented object; and
mapping the at least one indent node with the one or more defined features to generate a relationship between the one or more defined feature nodes and at least one indent node, wherein a defined feature node represents a defined feature in the knowledge base representation.
10. The system as claimed in claim 6 , wherein the processor is further configured to:
update the knowledge base representation of the object relationship data with new NN based on the one or more defined features automatically; and
transmit the data predicted and recognized from the selected NN to a decision-making module for further processing.
11. A non-transitory computer readable medium including instructions stored thereon that when processed by at least one processor cause a system for recognition of data to perform acts of:
receiving the one or more images from an image sensor coupled with the processor;
segmenting the one or more images to identify one or more segmented objects in the one or more images;
generating an object relationship data for each of the segmented objects, each of the segmented objects comprising at least one or more defined features and a confidence score;
determining a knowledge base representation of the object relationship data based on the one or more defined features of the one or more segmented objects, wherein the knowledge base representation comprises one or more defined feature nodes, one or more indent nodes and a plurality of links defining a relationship between the one or more defined feature nodes and at least one indent node;
training a Recurrent Neural Network (RNN) based on the knowledge base representation to generate a trained RNN; and
determining an appropriate Neural Network (NN) based on selection of the at least one indent node having optimum confidence score and the trained RNN for data prediction and recognition, wherein the instructions stored thereon causes the at least one processor to determine the appropriate NN by selecting the NN by the RNN based on at least one defined feature node linked with the indent node representing the selected NN.
12. The non-transitory computer readable medium as claimed in claim 11 , wherein the instructions stored thereon further causes the at least one processor to segment the one or more images by:
identifying one or more trained objects and text data in the one or more images;
generating at least one boundary for the one or more trained objects text data identified in the one or more images; and
cropping the identified trained objects and text data along the at least one boundary to determine the one or more segmented objects in the one or more images.
13. The non-transitory computer readable medium as claimed in claim 11 , wherein the instructions stored thereon further causes the at least one processor to generate the object relationship data by:
training a plurality of NNs with the one or more segmented objects;
identifying at least one defined feature for each of the one or more segmented objects; and
determining the confidence score for the one or more defined features of one or more segmented objects associated with the plurality of NNs.
14. The non-transitory computer readable medium as claimed in claim 11 , wherein the instructions stored thereon further causes the at least one processor to determine the knowledge base representation of the object relationship data by:
identifying the one or more indent nodes for each segmented object in the object relationship data, wherein an indent node represents a NN;
determining at least one indent node with optimum confidence score associated with each segmented object; and
mapping the at least one indent node with the one or more defined features to generate a relationship between the one or more defined feature nodes and at least one indent node, wherein a defined feature node represents a defined feature in the knowledge base representation.
15. The non-transitory computer readable medium as claimed in claim 11 , wherein the instructions stored thereon further causes the at least one processor to recognize data in one or more images by:
automatically updating the knowledge base representation of the object relationship data with new NN based on the one or more defined features; and
transmitting the data predicted and recognized from the selected NN to a decision-making module for further processing.