Thumbnail Crop Selection Using Supervised Learning Regression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for generating thumbnails often result in omitted important content, such as faces or key text, when cropping images to fit a fixed aspect ratio, leading to lower quality thumbnails, especially when automated processes are used, which can be time-consuming and costly to correct manually.
Innovation Solution
A processor-executed method using supervised learning to extract pixel-level features, generate unique bounding boxes, calculate region-level features, and learn a regression model to select optimal thumbnail crops, mimicking human editor cropping for improved quality and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If automated cropping processes are used to generate thumbnails, then productivity is improved, but manufacturing precision deteriorates
Solution Approach 1:
The system copies human editor behavior by training a machine learning model on examples of manually cropped thumbnails. The model learns to replicate the decision-making process of human editors in selecting optimal crop regions, thereby achieving high-quality thumbnails through automated processes without requiring actual human involvement in each case.
Solution Approach 2:
The system changes parameters by considering multiple candidate crop regions with different coordinates and aspect ratios, then uses a learned model to select the optimal parameters (crop location and size) that maximize thumbnail quality while maintaining the desired aspect ratio constraints.
2Manufacturing precision
If human editors perform cropping, then manufacturing precision is improved, but loss of time increases
Solution Approach 1:
The system performs preliminary action by pre-training a machine learning model on a dataset of manually cropped thumbnails before actual thumbnail generation. This preliminary training phase captures human editing expertise, allowing the model to automatically perform high-quality cropping without requiring human editors to spend time on each individual image during production.
Solution Approach 2:
The system enables self-service by allowing the machine learning model to autonomously select optimal crop regions without human intervention. Once trained, the model independently evaluates candidate crops and selects the best option based on learned patterns from human editor behavior, eliminating the need for human editors to manually crop each thumbnail.
3Ease of operation
If simple resizing is used to create thumbnails, then ease of operation is improved, but manufacturing precision deteriorates
Solution Approach 1:
The system applies segmentation by dividing the source image into multiple candidate crop regions and evaluating each segment independently. Instead of simply resizing the entire image, the system generates several possible crop variants, scores them using the learned model, and selects the best segment to ensure important content is preserved while maintaining the desired aspect ratio.
Data Source
AI summary
Software for supervised learning extracts a set of pixel-level features from each source image in collection of source images. Each of the source images is associated with a thumbnail created by an editor. The software also generates a collection of unique bounding boxes for each source image. And the software calculates a set of region-level features for each bounding box. Each region-level feature results from the aggregation of pixel values for one of the pixel-level features. The software learns a regression model, using the calculated region-level features and the thumbnail associated with the source image. Then the software chooses a thumbnail from a collection of unique bounding boxes in a new image, based on application of the regression model. The software uses a thumbnail received from an editor instead of the chosen thumbnail, if the chosen thumbnail is of insufficient quality as measured against a scoring threshold.