Optical Character Recognition Latency Reduction via Image Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing reading machines for visually impaired individuals face challenges in efficiently processing and reading text due to the compute-intensive nature of optical character recognition (OCR), leading to high latency and reduced user experience.
Innovation Solution
A portable reading device that employs cooperative processing between a handheld device and a computing system, utilizing optical character recognition and text-to-speech synthesis to quickly process recognized text by partitioning processing and using user-directed input to select sections of an image for processing, thereby reducing latency and improving efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the entire image is processed with optical character recognition, then complete text recognition is achieved, but processing time and latency increase significantly
Solution Approach 1:
The patent divides the image into multiple sections or regions, processing each section separately with optical character recognition. This segmentation allows the system to process portions of the image in parallel or sequentially with shorter wait times, reducing overall latency while maintaining complete text recognition across all sections.
2Measurement precision
If optical character recognition is performed on the entire image, then all text is recognized, but compute resources are overwhelmed and processing efficiency decreases
Solution Approach 1:
By segmenting the image into smaller sections, the computational load is distributed across multiple processing tasks. Each section requires fewer compute resources individually, allowing the system to maintain high text recognition accuracy while improving overall processing efficiency through parallel or staged execution.
Solution Approach 2:
The system processes sections of the image progressively rather than requiring complete image processing before producing any output. This partial action approach allows text recognition to begin on processed sections while other sections are still being processed, improving productivity without sacrificing final recognition accuracy.
3Reliability
If the system waits for complete image processing before providing results, then accurate text-to-speech conversion is achieved, but user experience deteriorates due to long waiting times
Solution Approach 1:
The system performs preliminary processing on sections of the image and prepares text-to-speech conversion for processed portions before the entire image is fully processed. This allows the system to provide preliminary accurate results to the user while continuing to process remaining sections, improving user experience without compromising final accuracy.
Solution Approach 2:
The text-to-speech synthesis operates continuously as sections are processed, rather than waiting for complete image processing. This continuous useful action provides users with ongoing accurate text conversion, maintaining reliability while significantly reducing perceived waiting time and improving ease of operation.
Data Source
AI summary
A portable reading device includes a computing device and a computer readable medium storing a computer program product to receive an image and select a section of the image to process. The product processes the section of the image with a first process and when the first process is finished processing the section of the image, process a result of the first process with a second process. While the second process is processing, repeats the first process on another section of the image.