Cross-Correlation Loss for Unsupervised 3D Object Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current 3D object detection technologies in advanced driver assistance systems (ADAS) face challenges in accurately estimating the location and classification of objects using diverse data modalities like images and point cloud data without labeled information, leading to suboptimal performance in unsupervised training scenarios.
Innovation Solution
A processor-implemented method that generates a cross-correlation loss by comparing feature vectors from neural networks trained on different data modalities, such as images and lidar point cloud data, to enhance the training of a neural network model for 3D object estimation, using techniques like dropout layers and random Gaussian noise for data augmentation, allowing for unsupervised training without label information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If unsupervised training is performed without label information, then the training complexity is reduced and ease of operation is improved, but the measurement precision of object detection deteriorates
Solution Approach 1:
The patent introduces a cross-correlation loss function as an intermediary mechanism that bridges the gap between unsupervised training simplicity and supervised detection accuracy. This loss function computes correlations between feature vectors from different modalities (e.g., image and point cloud) to provide implicit supervision signals, enabling the model to learn accurate object representations without requiring explicit label information during training
2Adaptability or versatility
If multiple data modalities are integrated, then the adaptability and robustness of the detection system are improved, but the device complexity increases
Solution Approach 1:
The patent merges multiple data modalities (such as image data and point cloud data) into a unified neural network training framework. By processing different modalities through separate neural networks and then integrating their feature vectors via cross-correlation computation, the system achieves multi-modality adaptability while managing complexity through modular architecture design
Solution Approach 2:
The patent creates a universal training framework that can handle multiple data modalities through a common cross-correlation loss mechanism. The same loss function and training procedure work across different modality combinations (image-point cloud, image-radar, etc.), providing multi-functionality without requiring separate specialized systems for each modality pair
Data Source
AI summary
A method and apparatus with object estimation model training is provided. The method include generating a cross-correlation loss based on a first feature vector, generated using an interim first neural network (NN) model provided an input based on first input data about a target object, and a second feature vector generated using a trained second neural network provided another input based on second input data about the target object; and generating a trained first NN model, including training the interim first NN model based on the cross-correlation loss.