Vehicle Image Classification System for Dealership Metadata

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current dealership websites fail to provide consumers with a complete understanding of specific vehicles on the lot, as they use generic 'stock images' or 'lot images' lacking metadata, which do not accurately represent the vehicle's features or views, making it inefficient for dealerships to manually determine and coordinate image and textual information.

Innovation Solution

An automated vehicle image classification system using machine learning techniques analyzes digital images to determine views and features, assigning classification IDs, and integrates with dealership websites to render accurate images and textual information, enabling consumers to view specific vehicle details.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of information

If dealerships use stock images or lot images without metadata, then the website is simpler to maintain, but consumers cannot obtain a complete understanding of specific vehicle features and views

Engineering Contradiction:

Improvevehicle feature and view informationVSAvoidimage metadata coordination system

Core Design Contradiction:

Loss of informationVSDevice complexity

Solution Approach 1:

The machine learning model automatically analyzes vehicle images and generates classification IDs describing views and features without requiring manual intervention from dealerships. The system self-services the metadata generation process, eliminating the need for human annotators to manually tag each image while maintaining comprehensive and accurate vehicle information.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent replaces the manual mechanical process of having photographers and staff manually tag and coordinate image metadata with an automated machine learning-based classification system. The ML model processes images algorithmically to generate structured metadata, substituting human labor with computational intelligence.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Productivity

If dealerships manually determine and coordinate image and textual information, then accuracy is improved, but time and resources required increase significantly

Engineering Contradiction:

Improvevehicle information management efficiencyVSAvoidtime to manage and display vehicle information

Core Design Contradiction:

ProductivityVSLoss of time

Solution Approach 1:

The machine learning model performs preliminary analysis of vehicle images to generate classification IDs before the images are needed for website display. This pre-processing action automates the information extraction and coordination process, eliminating the need for manual analysis at the time of vehicle listing or website updates.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system automatically manages the entire vehicle information coordination process including image analysis, feature identification, and metadata generation without requiring dealership staff intervention. The self-service capability dramatically reduces the time and resources needed for vehicle information management while maintaining high accuracy.

Inventive Principle:

Principle #25Self-service

3Measurement precision

If dealerships use third-party photographers to capture vehicle images, then image quality is improved, but the process becomes more complex and time-consuming

Engineering Contradiction:

Improveimage quality and feature accuracyVSAvoidimage capture and processing workflow

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent extracts the metadata generation function from the manual coordination process and integrates it directly into the image analysis workflow. By taking out the need for separate manual tagging operations and embedding automated ML-based classification within the image processing pipeline, the system maintains image quality while simplifying the overall workflow.

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The system merges the image capture process with the metadata generation process by using machine learning models to automatically analyze and classify images in real-time. This consolidation eliminates the need for separate manual coordination steps between photography and data entry operations, reducing workflow complexity while maintaining precision.

Inventive Principle:

Principle #5Merging (Combining)

Data Source

PatentUS11270168B1Method and system for vehicle image classification

Publication Date: 2022.03.08 AUTODATA SOLUTIONS INC A DE

AI summary

A method is disclosed that includes operations of receiving user input that includes a vehicle identifier (ID), responsive to receiving the user input, (1) determining a set of vehicle images corresponding to the vehicle ID, (2) a set of classification IDs corresponding to the set of vehicle images, wherein the set of classification IDs includes a classification ID for each vehicle image of the set of vehicle images, and (3) determining feature content corresponding to the set of classification IDs, generating an image-to-feature data map configured to associate the following (i) each vehicle image, (ii) one or more portions of the feature content, and (iii) one or more classification IDs, and transmitting the image-to-feature association to a logic module embedded in webpage code of a webpage, wherein the image-to-feature association includes instructions that, upon execution, cause a rendering of a first vehicle image and first feature content on the webpage.