Event Cameras offer appealing advantages, including power efficiency and ultra-low latency, driving forward advancements in edge applications. In order to leverage mature frame-based algorithms, most approaches typically compute dense, image-like representations from sparse, asynchronous events. However, they are often unable to capture comprehensive information or are computationally intensive, which hinders the edge deployment of event-based vision. Meanwhile, pillar-based paradigms have been proven to be efficient and well established for dense representations of sparse data. Hence, from a novel pillar-based perspective, we present EventPillars, an efficient, comprehensive framework for dense event representations. To summarize, it (i) incorporates the Temporal Event Range to describe an intact temporal distribution, (ii) Activates the Event Polarities to explicitly record the scene dynamics, (iii) enhances the target awareness by a spatial attention prior from Normalized Event Density, (iv) can be plug-and-played into different downstream tasks. Extensive experiments show that our EventPillars records a new state-of-the-art precision on object recognition and detection datasets with surprisingly 9.2× and 4.5× lower computation and storage consumption. This brings a new insight into dense event representations and is promising to boost the edge deployment of event-based vision.

Current LiDAR-Camera fusion methods for 3D object detection achieve considerable accuracy at the immense cost of computation and storage, posing challenges for the deployment at the edge. To address this issue, we propose a lightweight 3D object detection framework, namely TinyFusionDet. Specially, we put forward an ingenious Hybrid Scale Pillar Strategy in LiDAR point cloud feature extraction to efficiently improve the detection accuracy of small objects. Meanwhile, a low cost Cross-Modal Heatmap Attention module is presented to suppress background interference in image features for reducing false positives. Moreover, a Cross-Modal Feature Interaction module is designed to enhance the cross-modal information fusion among channels for further promoting the detection precision. Extensive experiments demonstrated that TinyFusionDet achieves competitive accuracy with the lowest memory consumption and inference latency, making it suitable for hardware constrained edge devices. Furthermore, TinyFusionDet is implemented on a customized FPGA-based prototype system, yielding a record high energy efficiency up to 114.97GOPS/W. To the best of our knowledge, this marks the first real-time LiDAR-Camera fusion detection framework for edge applications.

Look-up tables(LUTs)-based methods have recently shown enormous potential in image restoration tasks, which are capable of significantly accelerating the inference. However, the size of LUT exhibits exponential growth with the convolution kernel size, creating a storage bottleneck for its broader application on edge devices. Here, we address the storage explosion challenge to promote the capacity of mapping the complex CNN models by LUT. We introduce an innovative separable mapping strategy to achieve over $7\times$ storage reduction, transforming the storage from exponential dependence on kernel size to a linear relationship. Moreover, we design a dynamic discretization mechanism to decompose the activation and compress the quantization scale that further shrinks the LUT storage by $4.48\times$. As a result, the storage requirement of our proposed TinyLUT is around 4.1\% of MuLUT-SDY-X2 and amenable to on-chip cache, yielding competitive accuracy with over $5\times$ lower inference latency on Raspberry 4B than FSRCNN. Our proposed TinyLUT enables superior inference speed on edge devices with new state-of-the-art accuracy on both of image super-resolution and denoising, showcasing the potential of applying this method to various image restoration tasks at the edge. The codes are available at: https://github.com/Jonas-KD/TinyLUT.

In spite of achieving promising results in hyperspectral image (HSI) restoration, deep-learning-based methodologies still face the problem of spectral or spatial information loss due to neglecting the inner correlation of HSI. To address this issue, we propose an innovative deep recurrent convolution neural network (DnRCNN) model for HSI destriping. To the best of our knowledge, this is the first study on HSI destriping from the perspective of inner band and interband correlation explorations with the recurrent convolution neural network. In the novel DnRCNN, a selective recurrent memory unit (SRMU) is designed to respectively extract the correlative features involved in spectral and spatial domains. Moreover, an innovative recurrent fusion (RF) strategy incorporated with group concatenation is further proposed to remove strip noise and preserve scene details using the complementary features from SRMU. Experimental results on extensive HSI datasets validated that the proposed method achieves a new state-of-the-art (SOTA) HSI destriping performance.

The speckle interference seriously degrades the quality of synthetic aperture radar (SAR) image. The existing despeckling algorithms still struggle to remove noise and preserve details simultaneously. In order to enhance the noise suppression and detail restoration performance, this article specially presents a spatial and transform domain convolutional neural network (STD-CNN) model, which yields an integrated feature representation and learning framework for despeckling. In addition, an innovative feature refinement strategy is proposed to further reduce the detail loss by isolating detail features from noise features. Extensive experiments on synthetic and real SAR images demonstrate that the proposed method outperforms the existing SAR despeckling methods on both quantitative and qualitative assessments. With partial modification, the STD-CNN model can still be extended to other image restoration tasks.

This paper introduces deformable convolution in deep learning based joint denoising and demosaicing (JDD), which yields more adaptable representation and larger receptive fields in features extraction for a superior restoration performance. However, the deformable convolution generally leads to considerable computational load and irregular memory access bottleneck, limiting its extensive deployment on edge devices. To address this issue, we develop grouping strategy and assign independent offsets to each kernel group to reduce the computation latency while keeping the accuracy. Motivated by the exploration for aggregate distribution characteristics of deformable offsets, we present the offset sharing methodology to simplify the memory access complexity of deformable convolution. As for hardware acceleration, we specially design a novel deformable matrix multiplication workflow incorporated with a deformable memory mapping unit to boost the computational throughput. The verification experiments on FPGA demonstrate that the proposed deformable convolution based JDD can restore 4K Ultra High Definition (UHD) images at 70FPS and yields significant promotion in visual effect and objective quality assessment.