Abstract: Features for analyzing one or more images with image recognition to detect bounds to accurately convert physical text (e.g., pilot logbooks, aviation-related documents, journey logbooks, aircraft mechanics work orders, airframe logbooks, propeller logbooks, and engine logbooks) into digital format. Based on the predefined bounds, character recognition, e.g., OCR/ICR/ICR engine(s), may be used to convert each cell to a digital format. A decision engine determines, based on reinforcement learning data, the correct value of each cell based on a confidence of the character recognition. The decision engine may be a series of logical filtering, sorting, context rules (e.g., REGEX rules for specific columns) and suggestions based on the reinforcement learning model used to normalize the imperfections associated with the character recognition technologies. The final result is a serialized digital format that can be exported to an endpoint (e.g., spreadsheets, database, etc.).

Abstract: Features for analyzing one or more images with image recognition to detect bounds to accurately convert physical text (e.g., pilot logbooks, aviation-related documents, journey logbooks, aircraft mechanics work orders, airframe logbooks, propeller logbooks, and engine logbooks) into digital format. Based on the predefined bounds, character recognition, e.g., OCR/ICR/ICR engine(s), may be used to convert each cell to a digital format. A decision engine determines, based on reinforcement learning data, the correct value of each cell based on a confidence of the character recognition. The decision engine may be a series of logical filtering, sorting, context rules (e.g., REGEX rules for specific columns) and suggestions based on the reinforcement learning model used to normalize the imperfections associated with the character recognition technologies. The final result is a serialized digital format that can be exported to an endpoint (e.g., spreadsheets, database, etc.).