Image Redaction via Pose Landmarking and Skin Mask Blending

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current systems for nudity detection and redaction in images lack accuracy and adaptability, often failing to effectively identify and obscure inappropriate content, especially in diverse cultural contexts, and are not optimized for real-time processing in online applications such as virtual fitting rooms and social media sharing.

Innovation Solution

A system that employs machine learning algorithms for body pose landmarking and skin segmentation, generating polygon and skin masks to blend into a final mask layer for real-time redaction of inappropriate image portions, using configuration sets to adapt to cultural standards and specific application requirements.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional nudity detection systems are used, then processing speed may be maintained, but detection accuracy and cultural adaptability deteriorate

Engineering Contradiction:
Improvenudity detection accuracyVSAvoidcultural context adaptability
Core Design Contradiction:
Measurement precisionVSAdaptability or versatility

Solution Approach 1:

The detection system is segmented into multiple specialized modules: pose estimation module, skin region detection module, and cultural context adaptation module. Each module handles specific aspects of nudity detection independently, allowing high accuracy in each function while maintaining overall system adaptability to different cultural standards through configurable parameters.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system uses configurable parameter sets that can be adjusted based on cultural context. By changing detection thresholds, body part definitions, and skin tone ranges as parameters, the system maintains high detection accuracy across different cultural standards without requiring complete retraining of the detection models.

Inventive Principle:
Principle #35Parameter changes

2Measurement precision

If complex machine learning algorithms are implemented, then detection accuracy improves, but processing time increases

Engineering Contradiction:
Improvenudity detection accuracyVSAvoidreal-time processing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The system performs preliminary pose estimation and key point detection before detailed skin region analysis. By pre-identifying body contours and critical regions in a first pass, the system reduces the computational burden of subsequent detailed detection, achieving both high accuracy and real-time performance.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The processing pipeline is segmented into multiple stages with decreasing computational complexity. Early stages use lightweight models for coarse detection, while later stages apply more complex algorithms only to regions of interest identified in previous stages, maintaining real-time performance with high detection accuracy.

Inventive Principle:
Principle #1Segmentation

3Reliability

If comprehensive redaction is applied, then user privacy is protected, but image quality and usability deteriorate

Engineering Contradiction:
Improveuser privacy protectionVSAvoidimage quality
Core Design Contradiction:
ReliabilityVSLoss of information

Solution Approach 1:

The redaction system applies different levels of obscuration to different body regions based on their sensitivity. Critical private areas receive full redaction for privacy protection, while less sensitive regions use lighter obscuration methods, preserving overall image quality and usability while maintaining reliable privacy protection where needed.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The system applies redaction selectively only to detected inappropriate regions rather than the entire image. This partial action approach ensures privacy protection is applied exactly where needed, minimizing information loss in the rest of the image and maintaining overall image quality and usability.

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentUS20240428418A1System and method for detecting and redacting inappropriate images
Publication Date: 2024.12.26 WALMART APOLLO LLC
  • US20240428418A1 patent drawing
  • US20240428418A1 patent drawing
  • US20240428418A1 patent drawing

AI summary

A method can include capturing, via a camera, an image for upload. The method further can include determining a polygon mask layer having one or more polygon masks for a user body in the image, comprising: (a) determining, via a body pose landmarking algorithm, key joint points for the user body based on joint landmarks defined in a configuration set; and (b) determining the polygon masks based on the key joint points and one or more regions of interest defined in the configuration set. The method additionally can include determining a skin mask layer with a skin mask for the image. The method also can include blending the one or more polygon masks and the skin mask into a final mask layer comprising a final mask for the image. The method further can include applying the final mask to the image. Other embodiments are disclosed.