Related Experiment Video
Updated: May 5, 2026
08:25
Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment
Published on: May 7, 2019
9.1K
LGMMFusion: A LiDAR-guided multi-modal fusion framework for enhanced 3D object detection.
Haixing Cheng1, Chengyong Liu1, Wenzhe Gu1
1China Coal Energy Research Institute Co., Ltd., Xi'an, Shaanxi Province, China.
Plos One
|September 4, 2025
Summary
This study introduces LGMMfusion, a novel framework for autonomous driving perception. It enhances small object detection by fusing LiDAR and camera data earlier and more effectively, improving accuracy and robustness.
Area of Science:
- Computer Vision
- Autonomous Driving Systems
- Sensor Fusion
Background:
- Small object detection is crucial for autonomous driving safety.
- Current multi-modal fusion methods often process sensor data separately, limiting performance.
- Challenges include sparse LiDAR data and low-resolution image features.
Purpose of the Study:
- To propose a novel LiDAR-guided multi-modal fusion framework (LGMMfusion) for improved object detection.
- To enhance feature correlation between LiDAR and camera data before fusion.
- To specifically address the challenges of small object detection in autonomous driving.
Main Methods:
- Leveraging LiDAR depth information to guide the generation of image Bird's Eye View (BEV) features.
- Promoting spatial interaction between point clouds and pixels prior to BEV feature fusion.
- Employing multi-head multi-scale self-attention and adaptive cross-attention mechanisms for feature alignment.
- Fusing enhanced LiDAR BEV and image BEV features for the detection head.
Main Results:
- LGMMfusion achieved 71.1% NDS and 67.3% mAP on the nuScenes validation set.
- Demonstrated improved detection of small objects.
- Showcased enhanced detection accuracy for most object categories.
Conclusions:
- LGMMfusion effectively enhances multi-modal data fusion for object detection in autonomous driving.
- The proposed LiDAR-guided approach improves feature representation and detection performance, particularly for small objects.
- This framework offers a promising direction for robust perception systems.
Related Concept Videos
02:16Light Acquisition
8.0K
In order to produce glucose, plants need to capture sufficient light energy. Many modern plants have evolved leaves specialized for light acquisition. Leaves can be only millimeters in width or tens of meters wide, depending on the environment. Due to competition for sunlight, evolution has driven the evolution of increasingly larger leaves and taller plants, to avoid shading by their neighbors with contaminant elaboration of root architecture and mechanisms to transport water and nutrients.
8.0K
01:18Imaging Biological Samples with Optical Microscopy
9.2K
Optical microscopy uses optic principles to provide detailed images of samples. Antonie van Leeuwenhoek designed the first compound optical microscope in the 17th century to visualize blood cells, bacteria, and yeast cells. In 1830, Joseph Jackson Lister created an essentially modern light microscope. The 20th century saw the development of microscopes with enhanced magnification and resolution.
In optical microscopy, the specimen to be viewed is placed on a glass slide and clipped on the stage...
In optical microscopy, the specimen to be viewed is placed on a glass slide and clipped on the stage...
9.2K
01:30Three-Dimensional Force System
3.2K
In mechanical engineering, a three-dimensional force system is a system of forces acting in three dimensions, with forces applied along the x, y, and z coordinate axes. The three-dimensional force system is an important concept in mechanical engineering, as it allows engineers to understand and analyze the behavior of objects and structures in three dimensions. By understanding the forces acting on a system, engineers can design more efficient and effective mechanical systems that can withstand...
3.2K