Machine Learning Video Frame Selection Using Cropping Confidence
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Manual cropping of large volumes of raw images is tedious, burdensome, and inefficient, and the artistic and aesthetic judgement required presents challenges to automating the cropping process.
Innovation Solution
A machine learning (ML) predictor program is trained using a plurality of training raw images associated with sets of training master images indicating cropping characteristics, allowing the program to predict cropping characteristics for raw images and automate the cropping process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If manual cropping is performed on large volumes of raw images, then high-quality cropped images with artistic and aesthetic judgement can be obtained, but the process becomes tedious, burdensome, and inefficient
Solution Approach 1:
The system enables automated cropping by training the ML predictor program to independently perform cropping decisions without human intervention. The program learns from training data and automatically applies cropping characteristics to raw images, making the system self-sufficient for high-volume processing while maintaining quality standards.
Solution Approach 2:
The patent replaces the manual mechanical process of cropping with an automated ML-based system. The ML predictor program substitutes human operators, using learned patterns from training data to automatically determine cropping characteristics, thereby eliminating the tedious and inefficient manual process while preserving quality.
2Manufacturing precision
If manual cropping is performed to achieve artistic and aesthetic judgement, then cropping quality is maintained, but automation becomes challenging
Solution Approach 1:
The system performs preliminary action by training the ML predictor program in advance using training raw images and training master images. This pre-training phase enables the system to learn cropping characteristics and artistic judgment criteria before actual automated cropping operations, making automation feasible while maintaining quality.
Solution Approach 2:
The patent uses copying by creating training master images that represent desired cropping outcomes. These master images serve as templates or copies of ideal cropped results, which the ML predictor program learns to replicate automatically, thereby capturing artistic and aesthetic judgment in a form that can be systematically copied and applied.
3Extent of automation
If ML predictor program is trained with training data, then automated cropping capability is achieved, but training time and computational resources are required
Solution Approach 1:
The training process is performed as a preliminary action before deployment. By completing the training phase in advance using training raw images and training master images, the system establishes automated cropping capability upfront. This one-time preliminary investment enables subsequent rapid automated processing without recurring training delays.
Data Source
AI summary
Example systems and methods of selection of video frames using a machine learning (ML) predictor program are disclosed. The ML predictor program may generate predicted cropping boundaries for any given input image. Training raw images associated with respective sets of training master images indicative of cropping characteristics for the training raw image may be input to the ML predictor, and the ML predictor program trained to predict cropping boundaries for raw image based on expected cropping boundaries associated training master images. At runtime, the trained ML predictor program may be applied to a sequence of video image frames to determine for each respective video image frame a respective score corresponding to a highest statistical confidence associated with one or more subsets of cropping boundaries predicted for the respective video image frame. Information indicative of the respective video image frame having the highest score may be stored or recorded.