ROMA: Run-Time Object Detection To Maximize Real-Time Accuracy
This paper analyzes the effects of dynamically varying video contents and detection latency on the real-time detection accuracy of a detector and proposes a new run-time accuracy variation model, ROMA, based on the findings from the analysis. ROMA is designed to select an optimal detector out of a set of detectors in real time without label information to maximize real-time object detection accuracy. ROMA utilizing four YOLOv4 detectors on an NVIDIA Jetson Nano shows real-time accuracy improvements by 4 to 37% for a scenario of dynamically varying video contents and detection latency consisting of MOT17Det and MOT20Det datasets, compared to individual YOLOv4 detectors and two state-of-the-art runtime techniques.
PDF Abstract
Datasets
MOTChallenge
Results from the Paper
Submit
results from this paper
to get state-of-the-art GitHub badges and help the
community compare results to other papers.
Methods
1x1 Convolution •
Average Pooling •
Batch Normalization •
Bottom-up Path Augmentation •
Convolution •
Cosine Annealing •
CSPDarknet53 •
CutMix •
Darknet-53 •
DropBlock •
FPN •
Global Average Pooling •
Grid Sensitive •
k-Means Clustering •
Label Smoothing •
Logistic Regression •
Max Pooling •
Mish •
PAFPN •
ReLU •
Residual Connection •
Sigmoid Activation •
Softmax •
Softplus •
Spatial Attention Module •
Spatial Pyramid Pooling •
Tanh Activation •
YOLOv3 •
YOLOv4
