In this paper, we consider the problem of pedestrian detection in natural scenes. Intuitively, instances pedestrians with different spatial scales may exhibit dramatically features. Thus, large variance instance scales, which results undesirable intracategory features, severely hurt performance modern object methods. We argue that issue can be substantially alleviated by divide-and-conquer philosophy. Taking as an example, illustrate how leverage philosophy to develop a Scale-Aware Fast...

This paper proposes a human-aware deblurring model that disentangles the motion blur between foreground (FG) humans and background (BG). The proposed is based on triple-branch encoder-decoder architecture. first two branches are learned for sharpening FG BG details, respectively; while third one produces global, harmonious results by comprehensively fusing multi-scale information from domains. further endowed with supervised, attention mechanism in an end-to-end fashion. It learns soft mask...

In this paper, we present an effective and efficient face deblurring algorithm by exploiting semantic cues via deep convolutional neural networks (CNNs). As images are highly structured share several key components (e.g., eyes mouths), the information of a provides strong prior for restoration. such, propose to incorporate global priors as input impose local structure losses regularize output within multi-scale CNN. We train network with perceptual adversarial generate photo-realistic...

Automatic diabetic retinopathy (DR) lesions segmentation makes great sense of assisting ophthalmologists in diagnosis. Although many researches have been conducted on this task, most prior works paid too much attention to the designs networks instead considering pathological association for lesions. Through investigating pathogenic causes DR advance, we found that certain are closed specific vessels and present relative patterns each other. Motivated by observation, propose a relation...

In this work, we consider the problem of pedestrian detection in natural scenes. Intuitively, instances pedestrians with different spatial scales may exhibit dramatically features. Thus, large variance instance scales, which results undesirable intra-category features, severely hurt performance modern object methods. We argue that issue can be substantially alleviated by divide-and-conquer philosophy. Taking as an example, illustrate how leverage philosophy to develop a Scale-Aware Fast...

Modeling layout is an important first step for graphic design. Recently, methods generating layouts have progressed, particularly with Generative Adversarial Networks (GANs). However, the problem of specifying locations and sizes design elements usually involves constraints respect to element attributes, such as area, aspect ratio reading-order. Automating attribute conditional remains a complex unsolved problem. In this article, we introduce Attribute-conditioned Layout GAN incorporate...

Accurate and reliable perception systems are essential for autonomous driving robotics. To achieve this, 3D object detection with multi-sensors is necessary. Existing detectors have significantly improved accuracy by adopting a two-stage paradigm that relies solely on LiDAR point clouds proposal refinement. However, the sparsity of clouds, particularly faraway points, makes it difficult LiDAR-only refinement module to recognize locate objects accurately. address this issue, we propose novel...

The perception of drones, also known as Unmanned Aerial Vehicles (UAVs), particularly in infrared videos, is crucial for effective anti-UAV tasks. However, existing datasets UAV tracking have limitations terms target size and attribute distribution characteristics, which do not fully represent complex realistic scenes. To address this issue, we introduce a generalized benchmark called <bold xmlns:mml="http://www.w3.org/1998/Math/MathML"...

Layout is important for graphic design and scene generation. We propose a novel Generative Adversarial Network, called LayoutGAN, that synthesizes layouts by modeling geometric relations of different types 2D elements. The generator LayoutGAN takes as input set randomly-placed elements uses self-attention modules to refine their labels parameters jointly produce realistic layout. Accurate alignment critical good layouts. thus differentiable wireframe rendering layer maps the generated layout...

Layout is important for graphic design and scene generation. We propose a novel Generative Adversarial Network, called LayoutGAN, that synthesizes layouts by modeling geometric relations of different types 2D elements. The generator LayoutGAN takes as input set randomly-placed elements, represented vectors uses self-attention modules to refine their labels parameters jointly produce realistic layout. Accurate alignment critical good layouts. We, thus, differentiable wireframe rendering layer...

Liquid crystal tunable filters (LCTF) are extensively used in hyperspectral imaging systems to successively acquire different spectral components of scenes by adjusting the center wavelength filter. However, and spatial resolutions imager limited bandwidth LCTF, pitch dimension detector, respectively. This paper applies compressive sensing principles improve both LCTF-based system. An accurate transmission model LCTF is represent its bandpass filtering effects on spectra. In addition, a...

Discriminative Correlation Filter (DCF) based trackers are quite efficient in tracking objects by exploiting the circulant structure. The kernel trick further improves performance of such trackers. unwanted boundary effects, however, difficult to solve kernelized correlation models. In this paper, we propose a novel Constrained Multi-Kernel (CMKCF), which applies spatial constraints address drawback. We build multi-kernel models for multi-channel features with three different attributes, and...

Visual object tracking with semantic deep features has recently attracted much attention in computer vision. Especially, Siamese trackers, which aim to learn a decision making-based similarity evaluation, are widely utilized the community. However, online updating of fashion is still tricky issue due limitation, tradeoff between model adaption and degradation. To address such an issue, this article, we propose novel attentional transfer learning-based network (SiamATL), fully exploits...

Limited resolution is one of the most important factors hindering application remote sensing images (RSIs). Single-image super (SISR) a technique to improve spatial digital and has attracted attention many researchers. In recent years, with advancement deep learning (DL) frameworks, DL-based SISR models have been proposed achieved state-of-the-art performance; however, for RSIs use bicubic downsampler construct low-resolution (LR) high-resolution (HR) training pairs. Considering that quality...

Corrupted labels and class imbalance are commonly encountered in practically collected training data, which easily leads to over-fitting of deep neural networks (DNNs). Existing approaches alleviate these issues by adopting a sample re-weighting strategy, is re-weight designing weighting function. However, it only applicable for data containing either one type biases. In practice, however, biased samples with corrupted tailed classes co-exist data. How handle them simultaneously key but...

The world’s rapid industrialisation and population expansion have led to water pollution, causing significant disruption the activities of humans, animals, plants. Organic contamination content in is commonly evaluated by measuring chemical oxygen demand (COD). However, traditional COD detection methods often require additional reagents, resulting secondary extended time. In this study, we propose implement a reflective system that measures UV-Vis absorption spectra without contact...

Fine-grained geometry, obtained through the assimilation of localized point features, is crucial in realms object recognition and scene comprehension within cloud contexts. Traditional backbones predominantly utilize max pooling for amalgamation local a process that tends to overlook spatial interrelations among points, consequently leading potential loss fine-grained geometric details. To overcome this limitation, we introduce an innovative operation termed Position Adaptive Pooling...

White blood cells (WBC) play a significant role in human immune system, so WBC detection is meaningful work. In this paper, we propose novel framework which combines Fourier ptychographic microscopy (FPM) and SO-you only look once (YOLO) for WCB detection. FPM recently developed microscope technology can achieve high-resolution, wide-field, quantitative phase imaging at the same time without mechanical moving measurement. With FPM, get high resolution, wide field-of-view cell images one...