CLOCs: Camera-LiDAR Object Candidates Fusion for 3D Object Detection
Our most-cited work, pioneering a late-fusion approach for combining 2D and 3D detectors to achieve state-of-the-art 3D object detection.
Faculty Director
Daniel Morris, Ph.D.
Associate Professor
Biosystems & Agricultural Engineering
Electrical & Computer Engineering
Michigan State University
About
Research Vision
What is exciting right now is solving computer vision problems using machine learning and artificial intelligence techniques. Tasks that seemed way beyond reach a decade ago have been solved, but there are still many problems needing solutions.
Smart Agriculture, I believe, is the next big frontier for automation. We are exploring automated phenotyping and segmentation in cluttered scenes, as well as posture estimation of animals in depth images. I am actively involved in MSU NRT-IMPACTS, a program seeking to bridge the gap between computational and plant sciences through mentoring, teaching and research.
Automated Vehicles need vision to function safely and reliably. My students and I are finding ways to fuse video and lidar to obtain accurate, high-resolution depth maps that do not smear along the boundaries. We're also solving problems in radar-video fusion, tracking and object classification.
Expertise
Research Interests
3D Computer Vision
Developing algorithms for scene understanding from heterogeneous sensor data, enabling machines to perceive complex 3D environments.
Sensor Fusion
Fusing radar, LiDAR, and camera data to achieve superior accuracy for depth completion, object detection, and scene modeling.
Deep Learning
Applying convolutional neural networks and novel architectures to solve challenging detection, segmentation, and tracking problems.
Object Detection
Pioneering late-fusion approaches combining 2D and 3D detectors for robust real-time object detection in autonomous driving scenarios.
Precision Agriculture
Automated animal monitoring, posture estimation, crop phenotyping, and label-efficient learning for farm intelligence.
Tremor Analysis
Contactless tremor measurement via pose and optical flow fusion (POFF) for Parkinson's Disease patient assessment.
Featured Work
Key Publications
Highlights spanning top venues including CVPR, ICCV, IROS, and WACV.
RICCARDO: Radar Hit Prediction and Convolution for Camera-Radar 3D Object Detection
State-of-the-art probabilistic radar target modeling fused with camera data for next-generation autonomous driving perception.
DecoupledGaussian: Object-Scene Decoupling for Physics-Based Interaction
Object-scene decoupling method enabling physics-based interaction with 3D Gaussian scenes.
Full-Velocity Radar Returns by Radar-Camera Fusion
Depth Completion with Twin Surface Extrapolation at Occlusion Boundaries
Radar-Camera Pixel Depth Association for Depth Completion
RADIANT: Radar-Image Association Network for 3D Object Detection
Depth Coefficients for Depth Completion
DIAT: Depth-Infrared Image Annotation Transfer for Pig-Pose Detection
Teaching
Courses
ECE 201
Undergraduate
Circuits and Systems I
Circuits are fun and this course is a great way to learn the basics of analyzing linear circuits. Topics include Ohm's law, Kirchhoff's circuit laws, superposition, Thevenin and Norton equivalent circuits, Node Voltage Analysis, Mesh Current Analysis, as well as first and second order circuit responses.
ECE/CSE 434
Senior / Graduate
Autonomous Vehicles
We are on the cusp of a transportation revolution as our vehicles will soon take over the task of navigating us to where we want to go. This class covers sensor processing, object detection, planning, and control. Hands-on with ROS 2, Python, OpenCV, PyTorch, Turtlebots, Gazebo, and the CARLA driving simulator.
ROS 2
Python
PyTorch
OpenCV
CARLA
Gazebo
ECE 802
Graduate
Deep Learning Algorithms and Benchmarks
Created a new graduate-level course covering modern deep learning architectures, training techniques, and benchmark evaluation for computer vision tasks.
PyTorch
CNNs
Transformers
GANs
Beyond Academia
Ventures & Collaborations
Motion Grazer AI
Co-Founder · 2021 – Present
Bringing computer vision technologies to precision livestock farming with automated monitoring and health assessment tools.
PoliMOVE-MSU Racing
Collaborator · Indy Autonomous Challenge
Joint MSU – Politecnico di Milano team. First Place Passing Competition ($150K) and Second Place Timed Trials ($50K) at Indianapolis Motor Speedway, Sep 2024. Set the Autonomous Hill-Climb Time Record at Goodwood Festival of Speed, UK, Jul 2024.
2022 Innovation Award
MSU CANR Innovation of the Year
Novel sensor systems that monitor female swine, estimating body condition and lameness for proactive veterinary treatment. Team: M. Benjamin, D. Morris, S. Yik, M. Lavagnino.
MSU NRT-IMPACTS
Active Participant
A program seeking to bridge the gap between computational and plant sciences through mentoring, teaching and collaborative research.
Mentoring
Graduate Students
Current and former PhD and Masters students advised as committee chair.
Current Students
Hafez IrshaidPhD, CSE · Expected 2029
Mk BasharPhD, CSE · Expected 2027
Moe ElarbiMasters Student
Graduated
Yunfei LongPhD, ECE · 2025
Su PangPhD, ECE · 2022
Saif ImranPhD, ECE · 2022
Jiajia LiMasters, ECE · 2021
Steven YikMasters, ECE · 2020
Steven ChaoMasters, ECE · 2015
Service
Professional Activities
Editorial Boards
Associate Editor
Frontiers in Plant Science
2024 – Present
Editorial Board Member
Electronics Journal
2021 – Present
Associate Editor
Frontiers in Precision Livestock Farming
2022 – 2023
Professional Memberships
Member
IEEE
2003 – Present
Member
ASABE
2022 – Present