MME-YOLO: Multi-Sensor Multi-Level Enhanced YOLO for Robust Vehicle Detection in Traffic Surveillance
Open accessThe DOI matches its Crossref record; 51 of the 51 references it cites were found in a registry or an index; 10 of the 19 sources its claims cite were checked for support.
Abstract
As an effective means of solving collision problems caused by the limited perspective on board, the cooperative roadside system is gaining popularity. To improve the vehicle detection abilities in such online safety systems, in this paper, we propose a novel multi-sensor multi-level enhanced convolutional network model, called multi-sensor multi-level enhanced convolutional network architecture (MME-YOLO), with consideration of hybrid realistic scene of scales, illumination, and occlusion. MME-YOLO consists of two tightly coupled structures, i.e., the enhanced inference head and the LiDAR-Image composite module. More specifically, the enhanced inference head preliminarily equips the network with stronger inference abilities for redundant visual cues by attention-guided feature selection blocks and anchor-based/anchor-free ensemble head. Furthermore, the LiDAR-Image composite module cascades the multi-level feature maps from the LiDAR subnet to the image subnet, which strengthens the generalization of the detector in complex scenarios. Compared with YOLOv3, the enhanced inference head achieves a 5.83% and 4.88% mAP improvement on visual dataset LVSH and UA-DETRAC, respectively. Integrated with the composite module, the overall architecture gains 91.63% mAP in the collected Road-side Dataset. Experiments show that even under the abnormal lightings and the inconsistent scales at evening rush hours, the proposed MME-YOLO maintains reliable recognition accuracy and robust detection performance.
Overview
CoChat Summary
TL;DR
MME-YOLO is a multi-sensor vehicle detector that fuses LiDAR-derived depth/height/hue maps with image features at multiple levels, plus a cross-level attention and ensemble inference head. The paper reports improved vehicle-detection mAP over YOLOv3/YOLOv4 variants, especially for small vehicles and in adverse lighting/occlusion.
Key findings
- On the visual enhancement tests, E-YOLO achieved 72.48% mAP on LSVH and 92.99% mAP on UA-DETRAC.
- Compared with YOLOv3, the enhanced inference head improved mAP by 5.83% on LVSH and 4.88% on UA-DETRAC; the top-down attention structure alone improved overall detection by 2.47% on LSVH, with 5.36% improvement for small vehicles.
- On the collected roadside dataset of nearly 6K frames, adding LiDAR multi-level features made MME-YOLO improve over YOLOv3 by 7.46% mAP, and over the best YOLOv4-L by 1.66%.
- In the abstract, the full MME-YOLO system is reported to reach 91.63% mAP on the collected Road-side Dataset; the text also says the Car-category improvement grew from 0.80% to 7.37% when LiDAR was added.
Methods
- Two-part architecture: an E-YOLO visual enhancement module and an MME-YOLO multi-sensor composite backbone, both built on YOLOv3.
- Data/sample: evaluated on LSVH (14K images), UA-DETRAC (82K images), and a collected roadside dataset of about 6K frames from a HIKVISION DX-2XM6000 camera and Velodyne VLP-32 LiDAR.
- LiDAR processing projects point clouds to the image plane, interpolates sparse depth/height maps, extracts hue from images, and stacks depth + height + hue as LiDAR-subnet inputs; analysis is mainly mAP comparison, plus ablations on attention blocks and inference heads.
Limitations
- The paper states that MME-YOLO faces challenges from the effective distance of the LiDAR; it reports a minimum–maximum measurement distance of 93 m, beyond which cars are only occasionally captured.
- The text makes evident that the LiDAR-subnet design is still limited by sparse point-cloud resolution and inconsistent roadside distributions (the paper explicitly discusses these as sensor difficulties); this is a stated problem context, not a separate experimental limitation.
- The paper notes future work is needed for more weather conditions, more complex situations, time series, and other targets such as pedestrians and riders, indicating the current evaluation is limited in scope.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Figures





![Figure 6 Visualization of the multi-sensor multi-level composite module. The feature maps in the LiDAR subnet are added to the feature maps in the Image subnet in a recursive manner. This idea was originally published by CBNet [26]. We extend the composite methodology into the multi-sensor fusion field with a recursive multi-sensor multi-level composite structure.](https://figures.cochat.ai/paper-figures/5753ecab63d3440aa59ef7ee1e0fdb71/mme-yolo-multi-sensor-multi-level-enhanced-yolo-for-robust-vehicle-detection-in-figure-6.jpg)




Verification
Claims the paper makes
7 Valid · 6 Not valid · 8 Not checked · 49 Own results and interpretations
At least one cited source doesn’t support this claim or points to a different paper, and none confirmed it — see the verdicts below.
In the paper
“Since both these works only introduce LiDAR information as an additional input modality or cascading revision factors, the relation between LiDAR data and Image data at the semantic level is seriously neglected.”
§ MME-YOLO: Multi-Sensor Multi-Level Enhanced YOLO for Robust Vehicle Detection in Traffic Surveillance > 2. Related Work > 2.3. Deep-Leaning-Based Multi-Sensor Fusion Detectors from the Vehicle Perspective
What the sources say
- Not checked — no full text of the sourcecites [13] Chen, X., Ma, H., Wan, J., Li, B., and Xia, T. (2017, January 21–26). Multi-view 3d object detection network for autonomous driving. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, HI, USA. source
- Could not verifycites [14] Minemura, K., Liau, H., Monrroy, A., and Kato, S. (2018, January 21–23). LMNet: Real-time multiclass object detection on CPU using 3D LiDAR. Proceedings of the 3rd Asia-Pacific Conference on Intelligent Robot Systems (ACIRS), Singapore. sourceExperimental — accuracy not yet measured
Our copy of LMNet never makes the fusion critique the claim attributes to it: the Related Work (Sec. II) and Introduction (Sec. I) fault prior detectors (F-pointNet, VoxelNet, AVOD, DoBEM, MV3D) for pursuing accuracy while 'leav[ing] performance out of the scope' and for not being 'capable to perform on real-time scenarios' — a speed/real-time objection, not a semantic-level fusion objection. The only place image+LiDAR works are characterized together is the comparison table (Sec. IV, 'Network' — Input Data column listing 'Image and Lidar' for F-pointNet, AVOD, MV3D, DoBEM), which simply records their inputs and inference times and says nothing about neglecting the LiDAR–Image relation at the semantic level or about 'cascading revision factors'. Searched: 'multi-sensor fusion LiDAR image semantic level', 'AVOD MV3D additional input modality', 'drawbacks of combining RGB camera with LiDAR / why we do not use image data'; the paper's stated reason for avoiding RGB is low-light robustness (Abstract; Sec. II 'Manipulation of 3D coordinates'), so no passage in our copy either supports or contradicts the claimed criticism. (LLM, conf=0.70)
- Disputedcites [15] Qi, C.R., Liu, W., Wu, C., Su, H., and Guibas, L.J. (2018, January 18–23). Frustum pointnets for 3d object detection from rgb-d data. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Salt Lake City, UT, USA. sourceExperimental — accuracy not yet measured
The “3D Instance Segmentation PointNet” section explicitly says the method uses semantic categories from the 2D image detector to guide segmentation of the point cloud. The supplementary network-architecture section likewise says its segmentation and box-estimation networks leverage semantics predicted from RGB images. This contradicts the claim that the work seriously neglects the semantic relation between image and LiDAR data. (LLM, conf=0.88)
Checked by a reasoning model — may be wrong.
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“DT can effectively interpolate all locations in the map even original position is far from the nearest point.”
§ MME-YOLO: Multi-Sensor Multi-Level Enhanced YOLO for Robust Vehicle Detection in Traffic Surveillance > 3. The Proposed MME-YOLO Network > 3.3. LiDAR Point Image Generation
What the sources say
- Partly supportedcites [25] Lee, 1986, Generalized Delaunay triangulation for planar graphs, Discrete Comput. Geom. sourceExperimental — accuracy not yet measured
The Introduction (Section 1) of our copy does tie the Delaunay triangulation to interpolation, but only in a hedged and narrower form: linear interpolation is defined for a point p lying in the interior of a triangle of the grid, the performance 'depends heavily on the choice of the triangular grid', and the DT 'is believed to provide the best triangular grid for this purpose'. The citing sentence converts this belief-statement into a factual capability claim and extends it to 'all locations in the map' including positions 'far from the nearest point' — no passage in our copy (searched: interpolation/triangular grid quality; far from nearest data point, extrapolation outside the convex hull) addresses interpolation effectiveness at such distances, and the paper's own framing limits interpolation to points contained within the triangulation. So the DT-for-interpolation part is supported while the universality and far-from-nearest-point parts are not attributed to this source. (LLM, conf=0.72)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“Our works achieve better performance compared with recent robust detectors, such as [31] and [32], who also combine multi-level features to perform vehicle detection.”
§ MME-YOLO: Multi-Sensor Multi-Level Enhanced YOLO for Robust Vehicle Detection in Traffic Surveillance > 4. Experiments and Results > 4.2. Experiments for the Visual Enhancement
What the sources say
- Supportedcites [31] Zhang, F., Li, C., and Yang, F. (2019). Vehicle detection in urban traffic surveillance images based on convolutional neural networks with feature concatenation. Sensors, 19. source
The Introduction describes DP-SSD as a vehicle detector that concatenates feature maps from different layers, and Section 4.2 describes its vehicle detection as robust. The assertion that the citing authors’ own work performs better is their comparison, not a finding attributed to this source. (LLM, conf=0.90)
- Supportedcites [32] Zhang, 2019, CMNet: A connect-and-merge convolutional neural network for fast vehicle detection in urban traffic surveillance, IEEE Access source
The citation is attached to the identification of CMNet as a recent robust vehicle detector that combines multi-level features; the 'our works achieve better performance' comparison is the citing authors' own result, not an attribution to this source. Our copy supports the identification: the Introduction/Fig. 2 passage states that 'feature maps from lower layers can get higher localization precision, and upper layers can achieve better classification accuracy, thus we merge feature maps from MSPN with features of the same size from CMRN using concatenation' (a feature-pyramid-style multi-level fusion, Sec. III-B), and Sec. IV-C (Experiments on UA-DETRAC) describes the model as 'robust to detect vehicles with fine-grained categories for urban traffic surveillance', outperforming the compared detectors (Table 5, mAP 91.71%). (LLM, conf=0.85)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“When investigating the differences between E-YOLO and MME-YOLO, the relative mAP improvement of the Cars is boosted from 0.80% to 7.37% even under variable lighting conditions along with the occlusion”
§ MME-YOLO: Multi-Sensor Multi-Level Enhanced YOLO for Robust Vehicle Detection in Traffic Surveillance > 4. Experiments and Results > 4.3. Experiments for the Complex Scenarios in Urban Sections
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“As illustrated in Figure 10, the minimum–maximum measurement distance is 93 m.”
§ MME-YOLO: Multi-Sensor Multi-Level Enhanced YOLO for Robust Vehicle Detection in Traffic Surveillance > 4. Experiments and Results > 4.3. Experiments for the Complex Scenarios in Urban Sections
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
36 more claims
Create a free account to see all 70 claims and to verify the 5 references still open.
- Probably not a substantial claim. 18
- It may state two separate claims as one. 11
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ MME-YOLO: Multi-Sensor Multi-Level Enhanced YOLO for Robust Vehicle Detection in Traffic Surveillance > Abstract
“Compared with YOLOv3, the enhanced inference head achieves a 5.83% and 4.88% mAP improvement on visual dataset LVSH and UA-DETRAC, respectively.”
§ MME-YOLO: Multi-Sensor Multi-Level Enhanced YOLO for Robust Vehicle Detection in Traffic Surveillance > 4. Experiments and Results > 4.2. Experiments for the Visual Enhancement
“The mean average precision of different models on open dataset LSVH and UA-DETRAC are compared in Table 1.”
Citations
Cites
Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
Ren, S., He, K., Girshick, R., & Sun, J. (2017) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2016.2577031 · Cited by 73,472
Microsoft COCO: Common Objects in Context
Lin, T.-Y., Maire, M., Belongie, S., Hays, J., Perona, P., Ramanan, D., Dollár, P., & Zitnick, C. L. (2014) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-10602-1_48 · Cited by 45,460 · cited 1×
“In the first step, we adopt the original weight file of YOLOv3 as the base stone, which is trained on the COCO2017 [30] trainval dataset.”
MethodSqueeze-and-Excitation Networks
Hu, J., Shen, L., & Sun, G. (2018) · 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2018.00745 · Cited by 35,667
(2016) · Cited by 31,590
Feature Pyramid Networks for Object Detection
Lin, T.-Y., Dollar, P., Girshick, R., He, K., Hariharan, B., & Belongie, S. (2017) · 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2017.106 · Cited by 30,939 · cited 2×
“To alleviate both of the two problems, FPN [8] (the feature pyramid fusion network), which combines high-level semantic information with low-level visual information, is designed to reason objects on different scales.”
BackgroundFocal Loss for Dense Object Detection
Lin, T.-Y., Goyal, P., Girshick, R., He, K., & Dollar, P. (2017) · 2017 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2017.324 · Cited by 30,581 · cited 1×
“Specifically, lossf ocal refers to the focal loss [22] over vehicles’ category and lossDIOU refers to the DIOU [23] loss of vehicles’ bounding box on the lth level.”
BackgroundCBAM: Convolutional Block Attention Module
Woo, S., Park, J., Lee, J.-Y., & Kweon, I. S. (2018) · Lecture Notes in Computer Science · doi:10.1007/978-3-030-01234-2_1 · Cited by 25,853 · cited 1×
“The first kind is explicit attention, such as SENet [16], Residual Attention [17], CBAM [18], SKNet [19], ResneSt [20], etc.”
SSD: Single Shot MultiBox Detector
Liu, W., Anguelov, D., Erhan, D., Szegedy, C., Reed, S., Fu, C.-Y., & Berg, A. C. (2016) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-46448-0_2 · Cited by 22,128
YOLOv3: An Incremental Improvement
Redmon, J., & Farhadi, A. (2018) · arXiv · doi:10.48550/arxiv.1804.02767 · Cited by 17,141 · cited 16×
“The original backbones in each subnet are inherited from YOLOv3.”
MethodInfluentialSqueeze-and-Excitation Networks
Hu, J., Shen, L., Albanie, S., Sun, G., & Wu, E. (2020) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2019.2913372 · Cited by 12,157
Cited by
Computer Vision Applications in Intelligent Transportation Systems: A Survey
Dilek, E., & Dener, M. (2023) · Sensors · doi:10.3390/s23062938 · Cited by 171
Mittal, U., Chawla, P., & Tiwari, R. (2022) · Neural Computing and Applications · doi:10.1007/s00521-022-07940-9 · Cited by 159
Object Detection in Traffic Videos: A Survey
Ghahremannezhad, H., Shi, H., & Liu, C. (2023) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2023.3258683 · Cited by 141
“A number of studies tend to use different sensors such as Light Detection and Ranging (LiDAR) to benefit from additional depth information [186], [187], [188].”
BackgroundInfrastructure-Based Object Detection and Tracking for Cooperative Driving Automation: A Survey
Bai, Z., Wu, G., Qi, X., Liu, Y., Oguchi, K., & Barth, M. (2022) · IEEE Intelligent Vehicles Symposium · doi:10.1109/iv51971.2022.9827461 · Cited by 106
“1) Sensor System: Limited by the space, this paper does not provide a comprehensive overview for every related approaches, but most of the studies mentioned are based on a single sensor, while few of them are proposed for multi-sensor perception system [51], [72] in terms of roadside sensor-based”
BackgroundMethodA Multi-Scale Traffic Object Detection Algorithm for Road Scenes Based on Improved YOLOv5
Li, A., Sun, S., Zhang, Z., Feng, M., Wu, C., & Li, W. (2023) · Electronics · doi:10.3390/electronics12040878 · Cited by 82
Metadata
- Authors
- Zhu, J., Li, X., Jin, P., Xu, Q., Sun, Z., & Song, X.
- Affiliations
- Southeast University (CN)
- Traffic Management Research Institute (CN)
- Nanjing Xiaozhuang University (CN)
- Corresponding author
- Li, Xu
- Venue
- Sensors
- Publisher
- MDPI AG
- Published
- 2020-12-23
- Volume
- 21
- Issue
- 1
- Pages
- 27
- ISSN
- 1424-8220 (ISSN-L)
- PMC
- PMC7793071
- OpenAlex
- W3115103396
- Cited by
- 51
- References
- 51
- Citations by year
- 2026: 5 · 2025: 12 · 2024: 12 · 2023: 9 · 2022: 10 · 2021: 3
- FWCI
- 2.14
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- Sensors (publishedVersion, cc-by)
- Europe PMC (PubMed Central) (publishedVersion, other-oa)
- PubMed Central (publishedVersion, cc-by)
- PubMed (publishedVersion)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Funders
- Jiangsu Provincial Key Research and Development Program (BE2019106)
- National Natural Science Foundation of China (61801227)
- the Qing Lan Project of Jiangsu Province (QLGC2020)
- the Natural Science Foundation of the Jiangsu Higher Education Institutions of China (18KJB413007)
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
- Physical Sciences › Engineering › Automotive Engineering › Autonomous Vehicle Technology and Safety
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Video Surveillance and Tracking Methods
- Keywords
- vehicle detection, traffic monitoring, multi-sensor fusion, YOLO, occlusion handling, illumination robustness
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (qwen3.8-omni-flash and gpt-6-sol).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: MME-YOLO: Multi-Sensor Multi-Level Enhanced YOLO for Robust Vehicle Detection in Traffic Surveillance. https://app.cochat.ai/papers/10.3390/s21010027
- The paper
- Zhu, J., Li, X., Jin, P., Xu, Q., Sun, Z., & Song, X. (2020). MME-YOLO: Multi-Sensor Multi-Level Enhanced YOLO for Robust Vehicle Detection in Traffic Surveillance. Sensors. https://doi.org/10.3390/s21010027
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.