Video Object Detection

TDViT: Temporal Dilated Video Transformer for Dense Video Tasks
Guanxiong Sun, Yang Hua, Guosheng Hu, Neil Robertson
In ECCV, 2022
A transformer backbone designed for dense video tasks, e.g., video object detection, video instance segmentation.
TDViT: Temporal Dilated Video Transformer for Dense Video Tasks
Efficient One-stage Video Object Detection by Exploiting Temporal Consistency
Guanxiong Sun, Yang Hua, Guosheng Hu, Neil Robertson
In ECCV, 2022
We present a simple yet effecitve framework to address the computational bottlenecks when adapting SOTA video object detection methods to modern one-stage detectors.
Efficient One-stage Video Object Detection by Exploiting Temporal Consistency
MAMBA: Multi-level Aggregation via Memory Bank for Video Object Detection
Guanxiong Sun, Yang Hua, Guosheng Hu, Neil Robertson
In AAAI, 2021
We propose a novel memory bank to effectively model long-range temporal correlations between frames for video object detection. At the same time, our method can run in a very fast speed.
MAMBA: Multi-level Aggregation via Memory Bank for Video Object Detection