Information processing apparatus and non-transitory computer readable medium storing program
An information processing apparatus includes a processor configured to execute first acquisition processing for acquiring a first recognition result and a first recognition probability on target data from a first recognizer, execute second acquisition processing for acquiring a second recognition probability for the first recognition result on the target data from a second recognizer, and execute control for determining which of first processing and second processing with a necessary human workload greater than in the first processing is to be executed for the first recognition result based on the first recognition probability and the second recognition probability.
1. An information processing apparatus comprising:
a processor configured to
execute first acquisition processing for acquiring, from a first recognizer, a first recognition result and a first recognition probability on target data,
execute second acquisition processing for acquiring, from a second recognizer, a second recognition probability on the first recognition result acquired from the first recognizer, and
execute control for determining which of first processing and second processing with a necessary human workload greater than in the first processing is to be executed for the first recognition result based on the first recognition probability and the second recognition probability, wherein
the second recognizer obtains a plurality of recognition result candidates,
the processor acquires one of the plurality of recognition result candidates as the second recognition probability from the second recognizer, wherein the one of the plurality of recognition result candidates is the same as the first recognition result,
wherein, when the processor is configured to execute control for determining which of first processing and second processing is to be executed for the first recognition result, the first processing is executed in a case where the first recognition probability is higher than a first threshold value and the second recognition probability different from the first recognition probability is higher than a second threshold value, and the second processing is executed otherwise.
2. The information processing apparatus according to claim 1 ,
wherein, in the first acquisition processing, a plurality of candidates of the first recognition result on the target data and a first candidate recognition probability as a recognition probability for each candidate are acquired from the first recognizer,
in the second acquisition processing, for each candidate, a second candidate recognition probability as a recognition probability for the candidate on the target data is acquired from the second recognizer, and
when the processor is configured to execute control for determining which of first processing and second processing is to be executed for the first recognition result, in a case where there is a candidate, of which the first candidate recognition probability is higher than the first threshold value and the second candidate recognition probability is higher than the second threshold value, among the plurality of candidates, the first processing is executed on the candidate.
3. The information processing apparatus according to claim 1 ,
wherein the first recognizer recognizes a text string represented by a text string image as the target data and obtains the recognized text string as the first recognition result, and
the second recognizer recognizes a text represented by the text age for each text image included in the text string image and obtains, as the second recognition probability, a recognition probability for a recognition result representing a text corresponding to the text image in the text string represented by the first recognition result among a plurality of recognition results for the text image.
4. The information processing apparatus according to claim 3 ,
wherein, when the processor is configured to execute control for determining which of first processing and second processing is to be executed for the first recognition result, the first processing is executed in a case where the first recognition probability is higher than the first threshold value and the second recognition probability for each text in the text string represented by the first recognition result is higher than the second threshold value, and the second processing is executed otherwise.
5. The information processing apparatus according to claim 1 ,
wherein, in the first acquisition processing, a plurality of candidates of the first recognition result on the target data and a first candidate recognition probability as a recognition probability for each candidate are acquired from the first recognizer,
in the second acquisition processing, for each candidate, a second candidate recognition probability as a recognition probability for each text in the text string of the candidate on the target data is acquired from the second recognizer, and
when the processor is configured to execute control for determining which of first processing and second processing is to be executed for the first recognition result, in a case where there is a candidate, of which the first candidate recognition probability is higher than the first threshold value and the second candidate recognition probability for each text is higher than the second threshold value, among the plurality of candidates, the first processing is executed on the candidate.
6. The information processing apparatus according to claim 1 ,
wherein the first recognition result has the highest recognition probability among recognition results on the target data from the first recognizer.
7. A non-transitory computer readable medium storing a program that causes a computer to execute processing for
executing first acquisition processing for acquiring, from a first recognizer, a first recognition result and a first recognition probability on target data,
executing second acquisition processing for acquiring, from a second recognizer, a second recognition probability on the first recognition result acquired from the first recognizer, and
executing control for determining which of first processing and second processing with a necessary human workload greater than in the first processing is to be executed for the first recognition result based on the first recognition probability and the second recognition probability, wherein
the second recognizer obtains a plurality of recognition result candidates,
the computer acquires one of the plurality of recognition result candidates as the second recognition probability from the second recognizer, wherein the one of the plurality of recognition result candidates is the same as the first recognition result,
wherein, when the computer is configured to execute control for determining which of first processing and second processing is to be executed for the first recognition result, the first processing is executed in a case where the first recognition probability is higher than a first threshold value and the second recognition probability different from the first recognition probability is higher than a second threshold value, and the second processing is executed otherwise.