Neural Network Object Detection via Adaptive Image Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current electronic devices face challenges in efficiently identifying external objects using neural networks, particularly in segmenting images and distinguishing objects within complex scenes, which affects their performance in applications like autonomous driving and object recognition.
Innovation Solution
An electronic device equipped with a camera and processor that segments images into partial areas based on a field-of-view, identifies the presence of external objects using a neural network, and adjusts the segmentation gaps based on object detection, allowing for effective object identification and reduced processor usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the entire image is input to the neural network for object identification, then the identification accuracy is improved, but the processing time and computational resources increase
Solution Approach 1:
The image is divided into multiple regions of interest (ROIs) based on lane detection and vehicle position information. Instead of processing the entire image, the neural network only processes these segmented ROI portions, reducing computational load while maintaining identification accuracy for relevant objects
Solution Approach 2:
The patent extracts only the necessary portions of the image (ROIs) that contain potential target objects or are relevant to the detection task. By removing unnecessary image areas, the processing time is reduced while preserving the information needed for accurate object identification
2Productivity
If the image is segmented into smaller partial areas, then the processing efficiency is improved, but the object identification accuracy may deteriorate due to loss of contextual information
Solution Approach 1:
Before segmenting the image, the system performs preliminary processing to detect lanes and identify potential vehicle positions. This preliminary action guides the segmentation process to create ROIs that are more likely to contain relevant objects, ensuring that contextual information is preserved in the segmented regions
Solution Approach 2:
The system uses feedback from lane detection and initial object position estimation to adjust and optimize the ROI segmentation. The detected lane information and vehicle positions are fed back into the segmentation process to refine the boundaries of partial areas, ensuring that important contextual information is retained while maintaining processing efficiency
Data Source
AI summary
An electronic device may include a camera and a processor. The processor may be configured to: identify a width of a window to be used to segment the image, based on a field-of-view (FoV) of an image obtained through the camera, identify a height of the window based on a first area including a visual object corresponding to a reference surface, segment the first area to a plurality of partial areas, using the window, based on the width and the height, identify whether an external object is included in the first partial area, from a neural network to which a first partial area among the plurality of partial areas is inputted, and based on a gap identified based on whether the external object is included in the first partial area, obtain a second partial area separated from the first partial area within the image.