Semiconductor device, mobile apparatus, and method of controlling mobile apparatus

Solution Overview

Problem

Current image recognition systems in mobile apparatuses face increased processing load and power consumption when using high-resolution images or multiple images for object detection, which can degrade recognition accuracy.

Innovation Solution

A semiconductor device and mobile apparatus that includes an image detection unit to isolate object detection areas from captured images, an image recognition unit for processing these areas, and a control unit that adjusts movement speed and imaging intervals based on recognition probabilities, optimizing image recognition processing without excessive load.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If high-resolution captured images or multiple captured images are used for image recognition processing, then image recognition accuracy is improved, but processing load and power consumption increase

Engineering Contradiction:

Improveimage recognition accuracyVSAvoidpower consumption

Core Design Contradiction:

Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The captured image is divided into multiple regions of interest (ROIs) based on object detection results. Instead of performing image recognition processing on the entire high-resolution image, the system selectively processes only the detected object regions, significantly reducing the number of pixels subjected to recognition processing while maintaining recognition accuracy.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system extracts and processes only the necessary portions of the captured image (object regions) for recognition processing. By isolating and processing only the relevant object areas rather than the complete image, the computational load and power consumption are reduced while preserving the accuracy needed for effective recognition.

Inventive Principle:

Principle #2Taking out (Extraction)

2Measurement precision

If high-resolution captured images or multiple captured images are used for image recognition processing, then image recognition accuracy is improved, but processing load increases

Engineering Contradiction:

Improveimage recognition accuracyVSAvoidprocessing load

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The captured image is divided into multiple regions of interest (ROIs) based on object detection results. Instead of performing image recognition processing on the entire high-resolution image, the system selectively processes only the detected object regions, significantly reducing the number of pixels subjected to recognition processing while maintaining recognition accuracy.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system extracts and processes only the necessary portions of the captured image (object regions) for recognition processing. By isolating and processing only the relevant object areas rather than the complete image, the computational load and power consumption are reduced while preserving the accuracy needed for effective recognition.

Inventive Principle:

Principle #2Taking out (Extraction)

3Measurement precision

If multiple captured images are used for image recognition processing, then image recognition accuracy is improved, but the number of images to be processed increases

Engineering Contradiction:

Improveimage recognition accuracyVSAvoidprocessing efficiency

Core Design Contradiction:

Measurement precisionVSProductivity

Solution Approach 1:

The captured image is divided into multiple regions of interest (ROIs) based on object detection results. Instead of performing image recognition processing on the entire high-resolution image, the system selectively processes only the detected object regions, significantly reducing the number of pixels subjected to recognition processing while maintaining recognition accuracy.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system performs image recognition processing on a subset of the captured image data (only the object regions) rather than processing the complete image. This partial processing approach reduces the overall processing workload while still achieving the necessary recognition accuracy for the application.

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS11210530B2Semiconductor device, mobile apparatus, and method of controlling mobile apparatus

Publication Date: 2021.12.28 RENESAS ELECTRONICS CORP

AI summary

A semiconductor device for an apparatus having a movement drive unit and an image unit, includes an image detection unit, an image recognition unit and control unit. The image detection unit detects an object in a captured image and cuts out an image area including the object from the captured image as an object detection area image. The image recognition unit performs an image recognition processing for the object detection image area and output a recognition probability of the object. The control unit controls at least one of a moving speed of the movement drive unit and an imaging interval of the image unit based on the recognition probability.