# MAMBA: Multi-level Aggregation via Memory Bank for Video Object Detection

arXiv: https://arxiv.org/abs/2401.09923 · HTML: https://mlsanity.com/paper/2401.09923/

## Key facts

- Cited by 13 papers in the map (data through 14 Oct 2026).
- Placed in Point clouds, detection and remote sensing, one of 10 directions in State-space sequence models (S4, Mamba).
- Builds on 28 of its 32 references that are in the map.
- First posted on arXiv 18 Jan 2024.

## Placed in

- State-space sequence models (S4, Mamba) / [Point clouds, detection and remote sensing](https://mlsanity.com/ssm/point-clouds-detection-and-remote-sensing/)

## Builds on (most cited first, 28 of 28)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Very Deep Convolutional Networks for Large-Scale Image Recognition](https://mlsanity.com/paper/1409.1556/) (2014)
- [Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks](https://mlsanity.com/paper/1506.01497/) (2015)
- [ImageNet Large Scale Visual Recognition Challenge](https://mlsanity.com/paper/1409.0575/) (2014)
- [You Only Look Once: Unified, Real-Time Object Detection](https://mlsanity.com/paper/1506.02640/) (2015)
- [Going Deeper with Convolutions](https://mlsanity.com/paper/1409.4842/) (2014)
- [Fast R-CNN](https://mlsanity.com/paper/1504.08083/) (2015)
- [Aggregated Residual Transformations for Deep Neural Networks](https://mlsanity.com/paper/1611.05431/) (2016)
- [FCOS: Fully Convolutional One-Stage Object Detection](https://mlsanity.com/paper/1904.01355/) (2019)
- [YOLO9000: Better, Faster, Stronger](https://mlsanity.com/paper/1612.08242/) (2016)
- [FlowNet: Learning Optical Flow with Convolutional Networks](https://mlsanity.com/paper/1504.06852/) (2015)
- [FlowNet 2.0: Evolution of Optical Flow Estimation with Deep Networks](https://mlsanity.com/paper/1612.01925/) (2016)
- [R-FCN: Object Detection via Region-based Fully Convolutional Networks](https://mlsanity.com/paper/1605.06409/) (2016)
- [Relation Networks for Object Detection](https://mlsanity.com/paper/1711.11575/) (2017)
- [Deep Feature Flow for Video Recognition](https://mlsanity.com/paper/1611.07715/) (2016)
- [Flow-Guided Feature Aggregation for Video Object Detection](https://mlsanity.com/paper/1703.10025/) (2017)
- [Detect to Track and Track to Detect](https://mlsanity.com/paper/1710.03958/) (2017)
- [Memory Enhanced Global-Local Aggregation for Video Object Detection](https://mlsanity.com/paper/2003.12063/) (2020)
- [Sequence Level Semantics Aggregation for Video Object Detection](https://mlsanity.com/paper/1907.06390/) (2019)
- [Relation Distillation Networks for Video Object Detection](https://mlsanity.com/paper/1908.09511/) (2019)
- [Seq-NMS for Video Object Detection](https://mlsanity.com/paper/1602.08465/) (2016)
- [T-CNN: Tubelets with Convolutional Neural Networks for Object Detection from Videos](https://mlsanity.com/paper/1604.02532/) (2016)
- [Object Detection in Video with Spatiotemporal Sampling Networks](https://mlsanity.com/paper/1803.05549/) (2018)
- [Progressive Sparse Local Attention for Video object detection](https://mlsanity.com/paper/1903.09126/) (2019)
- [Object Detection in Videos with Tubelet Proposal Networks](https://mlsanity.com/paper/1702.06355/) (2017)
- [Towards High Performance Video Object Detection for Mobiles](https://mlsanity.com/paper/1804.05830/) (2018)
- [Object Detection in Video with Spatial-temporal Context Aggregation](https://mlsanity.com/paper/1907.04988/) (2019)

## Built on by (most cited first, 13 of 13)

- [Historical Test-time Prompt Tuning for Vision Foundation Models](https://mlsanity.com/paper/2410.20346/) (2024)
- [GeoPix: Multi-Modal Large Language Model for Pixel-level Image Understanding in Remote Sensing](https://mlsanity.com/paper/2501.06828/) (2025)
- [ColorMNet: A Memory-based Deep Spatial-Temporal Feature Propagation Network for Video Colorization](https://mlsanity.com/paper/2404.06251/) (2024)
- [Spatio-temporal Prompting Network for Robust Video Feature Extraction](https://mlsanity.com/paper/2402.02574/) (2024)
- [Efficient One-stage Video Object Detection by Exploiting Temporal Consistency](https://mlsanity.com/paper/2402.09241/) (2024)
- [Practical Video Object Detection via Feature Selection and Aggregation](https://mlsanity.com/paper/2407.19650/) (2024)
- [Mixture of Scale Experts for Alignment-free RGBT Video Object Detection and A Unified Benchmark](https://mlsanity.com/paper/2410.12143/) (2024)
- [Hierarchical Point-Patch Fusion with Adaptive Patch Codebook for 3D Shape Anomaly Detection](https://mlsanity.com/paper/2604.03972/) (2026)
- [A Comprehensive Survey on Video Scene Parsing:Advances, Challenges, and Prospects](https://mlsanity.com/paper/2506.13552/) (2025)
- [Beyond Boxes: Mask-Guided Spatio-Temporal Feature Aggregation for Video Object Detection](https://mlsanity.com/paper/2412.04915/) (2024)
- [TGBFormer: Transformer-GraphFormer Blender Network for Video Object Detection](https://mlsanity.com/paper/2503.13903/) (2025)
- [TDViT: Temporal Dilated Video Transformer for Dense Video Tasks](https://mlsanity.com/paper/2402.09257/) (2024)
- [Can Visual Mamba Improve AI-Generated Image Detection? An In-Depth Investigation](https://mlsanity.com/paper/2605.14799/) (2026)

## Abstract (arXiv)

State-of-the-art video object detection methods maintain a memory structure, either a sliding window or a memory queue, to enhance the current frame using attention mechanisms. However, we argue that these memory structures are not efficient or sufficient because of two implied operations: (1) concatenating all features in memory for enhancement, leading to a heavy computational cost; (2) frame-wise memory updating, preventing the memory from capturing more temporal information. In this paper, we propose a multi-level aggregation architecture via memory bank called MAMBA. Specifically, our memory bank employs two novel operations to eliminate the disadvantages of existing methods: (1) light-weight key-set construction which can significantly reduce the computational cost; (2) fine-grained feature-wise updating strategy which enables our method to utilize knowledge from the whole video. To better enhance features from complementary levels, i.e., feature maps and proposals, we further propose a generalized enhancement operation (GEO) to aggregate multi-level features in a unified manner. We conduct extensive evaluations on the challenging ImageNetVID dataset. Compared with existing state-of-the-art methods, our method achieves superior performance in terms of both speed and accuracy. More remarkably, MAMBA achieves mAP of 83.7/84.6% at 12.6/9.1 FPS with ResNet-101. Code is available at https://github.com/guanxiongsun/vfe.pytorch.