Sparse Transformer for Ultra-sparse Sampled Video Compressive Sensing

• URL: http://arxiv.org/abs/2509.08228v1
• Date: Wed, 10 Sep 2025 02:03:06 GMT
• Title: Sparse Transformer for Ultra-sparse Sampled Video Compressive Sensing
• Abstract summary: Digital cameras consume 0.1 microjoule per pixel to capture and encode video, resulting in a power usage of 20W for a 4K sensor operating at 30 fps.<n>We propose Ultra-Sparse Sampling (USS) regime, where at each spatial location, only one sub-frame is set to 1.<n>We then build a Digital Micro-mirror Device (DMD) encoding system to verify the effectiveness of our USS strategy.
• Score: 23.788589833081957
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Digital cameras consume ~0.1 microjoule per pixel to capture and encode video, resulting in a power usage of ~20W for a 4K sensor operating at 30 fps. Imagining gigapixel cameras operating at 100-1000 fps, the current processing model is unsustainable. To address this, physical layer compressive measurement has been proposed to reduce power consumption per pixel by 10-100X. Video Snapshot Compressive Imaging (SCI) introduces high frequency modulation in the optical sensor layer to increase effective frame rate. A commonly used sampling strategy of video SCI is Random Sampling (RS) where each mask element value is randomly set to be 0 or 1. Similarly, image inpainting (I2P) has demonstrated that images can be recovered from a fraction of the image pixels. Inspired by I2P, we propose Ultra-Sparse Sampling (USS) regime, where at each spatial location, only one sub-frame is set to 1 and all others are set to 0. We then build a Digital Micro-mirror Device (DMD) encoding system to verify the effectiveness of our USS strategy. Ideally, we can decompose the USS measurement into sub-measurements for which we can utilize I2P algorithms to recover high-speed frames. However, due to the mismatch between the DMD and CCD, the USS measurement cannot be perfectly decomposed. To this end, we propose BSTFormer, a sparse TransFormer that utilizes local Block attention, global Sparse attention, and global Temporal attention to exploit the sparsity of the USS measurement. Extensive results on both simulated and real-world data show that our method significantly outperforms all previous state-of-the-art algorithms. Additionally, an essential advantage of the USS strategy is its higher dynamic range than that of the RS strategy. Finally, from the application perspective, the USS strategy is a good choice to implement a complete video SCI system on chip due to its fixed exposure time.

Related papers

• bit2bit: 1-bit quanta video reconstruction via self-supervised photon prediction [57.199618102578576]
  We propose bit2bit, a new method for reconstructing high-quality image stacks at original resolution from sparse binary quantatemporal image data. Inspired by recent work on Poisson denoising, we developed an algorithm that creates a dense image sequence from sparse binary photon data. We present a novel dataset containing a wide range of real SPAD high-speed videos under various challenging imaging conditions.
  arXiv  Detail & Related papers  (2024-10-30T17:30:35Z)
• Deep Optics for Video Snapshot Compressive Imaging [10.830072985735175]
  Video snapshot imaging (SCI) aims to capture a sequence of video frames with only a single shot of a 2D detector. This paper presents a framework to jointly optimize masks and a reconstruction network. We believe this is a milestone for real-world video SCI.
  arXiv  Detail & Related papers  (2024-04-08T08:04:44Z)
• Privacy-Enhancing Optical Embeddings for Lensless Classification [17.169529483306103]
  We build a prototype from a low-cost LCD and Raspberry Pi components, for a total cost of around 100 USD. We show how the proposed system, through variable mask patterns, can thwart adversaries that attempt to invert the system.
  arXiv  Detail & Related papers  (2022-11-23T11:12:21Z)
• FOVEA: Foveated Image Magnification for Autonomous Navigation [53.69803081925454]
  We propose an attentional approach that elastically magnifies certain regions while maintaining a small input canvas. Our proposed method boosts the detection AP over standard Faster R-CNN, with and without finetuning. On the autonomous driving datasets Argoverse-HD and BDD100K, we show our proposed method boosts the detection AP over standard Faster R-CNN, with and without finetuning.
  arXiv  Detail & Related papers  (2021-08-27T03:07:55Z)
• TUM-VIE: The TUM Stereo Visual-Inertial Event Dataset [50.8779574716494]
  Event cameras are bio-inspired vision sensors which measure per pixel brightness changes. They offer numerous benefits over traditional, frame-based cameras, including low latency, high dynamic range, high temporal resolution and low power consumption. To foster the development of 3D perception and navigation algorithms with event cameras, we present the TUM-VIE dataset.
  arXiv  Detail & Related papers  (2021-08-16T19:53:56Z)
• 10-mega pixel snapshot compressive imaging with a hybrid coded aperture [48.95666098332693]
  High resolution images are widely used in our daily life, whereas high-speed video capture is challenging due to the low frame rate of cameras working at the high resolution mode. snapshot imaging (SCI) was proposed as a solution to the low throughput of existing imaging systems.
  arXiv  Detail & Related papers  (2021-06-30T01:09:24Z)
• Time-Multiplexed Coded Aperture Imaging: Learned Coded Aperture and Pixel Exposures for Compressive Imaging Systems [56.154190098338965]
  We show that our proposed time multiplexed coded aperture (TMCA) can be optimized end-to-end. TMCA induces better coded snapshots enabling superior reconstructions in two different applications: compressive light field imaging and hyperspectral imaging. This codification outperforms the state-of-the-art compressive imaging systems by more than 4dB in those applications.
  arXiv  Detail & Related papers  (2021-04-06T22:42:34Z)
• Memory-Efficient Network for Large-scale Video Compressive Sensing [21.040260603729227]
  Video snapshot imaging (SCI) captures a sequence of video frames in a single shot using a 2D detector. In this paper, we develop a memory-efficient network for large-scale video SCI based on multi-group reversible 3D convolutional neural networks.
  arXiv  Detail & Related papers  (2021-03-04T15:14:58Z)
• Plug-and-Play Algorithms for Video Snapshot Compressive Imaging [41.818167109996885]
  We consider the reconstruction problem of snapshot video imaging (SCI) using a low-speed 2D sensor (detector) The underlying principle SCI is to modulate frames with different masks and then encoded frames are integrated into a snapshot on the sensor. Applying SCI to largescale problems (HD or UHD videos) in our daily life is still challenging one bottlenecks lies in the reconstruction algorithm.
  arXiv  Detail & Related papers  (2021-01-13T00:51:49Z)
• StrObe: Streaming Object Detection from LiDAR Packets [73.27333924964306]
  Rolling shutter LiDARs emitted as a stream of packets, each covering a sector of the 360deg coverage. Modern perception algorithms wait for the full sweep to be built before processing the data, which introduces an additional latency. In this paper we propose StrObe, a novel approach that minimizes latency by ingesting LiDAR packets and emitting a stream of detections without waiting for the full sweep to be built.
  arXiv  Detail & Related papers  (2020-11-12T14:57:44Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.