Compressed Image Acquisition for Low-Power Scene Recognition

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing image acquisition devices face challenges in achieving compactness and miniaturization due to high power consumption and resource requirements for image processing and transmission, which hinder lightweight and portable design.

Innovation Solution

Implementing image compression and low-power communication technologies to reduce data volume and power consumption, allowing for efficient image recognition through a pre-bound image recognition device using a higher compression ratio and low-bandwidth communication.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If image data is transmitted without compression, then image recognition accuracy is maintained, but power consumption and data transmission volume increase

Engineering Contradiction:
Improveimage recognition accuracyVSAvoidpower consumption
Core Design Contradiction:
Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The patent applies preliminary action by performing image compression before transmission. The image acquisition device compresses the captured image data into a smaller format (e.g., converting full-resolution images to thumbnail versions) before sending to the recognition device. This pre-processing reduces the data volume that needs to be transmitted and processed, thereby lowering power consumption while still enabling effective recognition through subsequent upscaling or feature extraction techniques.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent segments the image processing workflow into two distinct stages: (1) initial compression and transmission of reduced-size images from the acquisition device, and (2) secondary processing and recognition at the recognition device. This segmentation allows each component to perform optimized operations - the acquisition device focuses on minimal viable data capture, while the recognition device handles the computationally intensive recognition tasks, thereby distributing and reducing overall power consumption.

Inventive Principle:
Principle #1Segmentation

2Use of energy by moving object

If image data is compressed at high compression ratio, then power consumption and device size are reduced, but image quality and recognition accuracy may deteriorate

Engineering Contradiction:
Improvepower consumptionVSAvoidimage recognition accuracy
Core Design Contradiction:
Use of energy by moving objectVSMeasurement precision

Solution Approach 1:

The system performs preliminary compression at the image acquisition device using optimized algorithms that preserve critical features while reducing data volume. The compression is designed to maintain sufficient image quality for recognition purposes by retaining key edges, contours, and distinctive features that are essential for accurate classification, while discarding redundant detail information.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent dynamically adjusts compression parameters based on the specific application requirements and available resources. By changing parameters such as compression ratio, resolution levels, and feature retention thresholds, the system optimizes the balance between power consumption and recognition accuracy for different scenarios, allowing flexible adaptation to various operational conditions.

Inventive Principle:
Principle #35Parameter changes

3Measurement precision

If full-resolution image data is transmitted, then recognition accuracy is maximized, but data transmission volume and processing requirements increase

Engineering Contradiction:
Improvescene recognition accuracyVSAvoiddata transmission volume
Core Design Contradiction:
Measurement precisionVSQuantity of substance

Solution Approach 1:

The image acquisition device performs preliminary compression to generate a condensed version of the image data before transmission. This pre-compression step significantly reduces the data volume that needs to be transmitted over the communication interface, lowering bandwidth requirements and transmission time while still preserving sufficient information for accurate scene recognition through subsequent processing at the recognition device.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent introduces an intermediary compression layer between the image capture sensor and the recognition algorithm. This intermediary processing stage transforms the raw high-resolution image data into a compressed representation that serves as an efficient intermediary format - small enough for low-bandwidth transmission but rich enough in preserved features to enable accurate recognition through intelligent algorithms at the receiving end.

Inventive Principle:
Principle #24Intermediary (Mediator)

4Productivity

If image sensor operates continuously, then image acquisition capability is maintained, but power consumption increases

Engineering Contradiction:
Improveimage acquisition capabilityVSAvoidpower consumption
Core Design Contradiction:
ProductivityVSUse of energy by moving object

Solution Approach 1:

The image sensor operates periodically rather than continuously, activating only when needed for capturing images based on trigger events, time intervals, or motion detection. This periodic operation mode allows the sensor to remain in a low-power state during non-acquisition periods while maintaining the ability to quickly capture images when required, thereby significantly reducing overall power consumption while preserving image acquisition capability.

Inventive Principle:
Principle #19Periodic action

Solution Approach 2:

The system dynamically adjusts the operational state of the image sensor based on real-time conditions and requirements. The sensor can switch between active capture mode and low-power standby mode, with transition triggers determined by application needs, motion detection, or scheduled intervals. This dynamic operation optimizes the balance between maintaining image acquisition readiness and minimizing power consumption during periods when imaging is not required.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP4704433A1Method for image acquisition and recognition, apparatus, and non-transitory computer-readable storage medium
Publication Date: 2026.03.04 BEIJING XIAOMI MOBILE SOFTWARE CO LTD
  • EP4704433A1 patent drawingFigure 1
  • EP4704433A1 patent drawingFigure 2~4
  • EP4704433A1 patent drawingFigure 5~6

AI summary

The present disclosure relates to a method for image acquisition and image recognition, an apparatus, an electronic device and a storage medium, and relates to the field of image processing technology. The method for image acquisition includes: acquiring first image data of a target scene; obtaining second image data by performing image compression on the first image data; and sending the second image data to a pre-bound image recognition device to cause the image recognition device to perform image recognition on the second image data through an image recognition model.