Enhanced Automotive Object Detection via RGB-D Fusion in a DiffusionDet Framework

TL;DR

This paper introduces a novel RGB-D fusion framework based on DiffusionDet for autonomous driving, significantly improving automotive object detection accuracy, especially for small objects, by combining camera and depth sensor data.

Contribution

It presents a new diffusion-based object detection framework that fuses RGB and depth data, enhancing detection accuracy over existing methods.

Findings

Achieved a 2.3 AP gain on the KITTI dataset.

Enhanced detection of small automotive objects.

Demonstrated effective RGB-D feature fusion in diffusion models.

Abstract

Vision-based autonomous driving requires reliable and efficient object detection. This work proposes a DiffusionDet-based framework that exploits data fusion from the monocular camera and depth sensor to provide the RGB and depth (RGB-D) data. Within this framework, ground truth bounding boxes are randomly reshaped as part of the training phase, allowing the model to learn the reverse diffusion process of noise addition. The system methodically enhances a randomly generated set of boxes at the inference stage, guiding them toward accurate final detections. By integrating the textural and color features from RGB images with the spatial depth information from the LiDAR sensors, the proposed framework employs a feature fusion that substantially enhances object detection of automotive targets. The $2.3$ AP gain in detecting automotive targets is achieved through comprehensive experiments using the KITTI dataset. Specifically, the improved performance of the proposed approach in detecting small objects is demonstrated.