Comparison Results: SSH-YOLO Model Performance on Road Scene Datasets

Tabular results from an experiment comparing the SSH-YOLO object detection model against YOLOv8n on several datasets. The data includes mean average precision (mAP) scores for models tested on the self-built RoadScene-Complex dataset and public datasets BDD100K, KITTI, COCO, and CityPersons. The dataset was authored by Tenglong Ma and last updated on April 15, 2026.

Use Cases

• Benchmarking object detection model performance based on reported [email protected] metrics.
• Comparing model accuracy for small targets (<32x32 pixels) based on results from the COCO small target subset.
• Evaluating model robustness in occluded scenes based on results from the CityPersons dataset.
• Assessing the trade-off between model accuracy and inference speed based on the reported 60 FPS metric.

Strengths

• Provides specific [email protected] performance gains (e.g., 12.4% higher than YOLOv8n on RoadScene-Complex).
• Reports model performance across multiple established public datasets (BDD100K, KITTI, COCO, CityPersons).
• Includes a concrete inference speed metric of 60 FPS.
• Released under a permissive CC-BY-4.0 license.

Limitations

• Row count and specific column-level documentation are absent; field semantics must be inferred after download.
• The dataset is very small (5.5 KB), suggesting it contains only summary results, not raw training or validation data.
• Data may reflect bias inherent to the specific road scene datasets used for evaluation.

Provenance

Source

Tenglong Ma via figshare.

Collection Method

Likely contains experimental results from model training and evaluation.

Freshness

Last updated 2026-04-15 18:05:43; freshness should be verified.