Image similarity based on Discrete Wavelet Transform for robots with low-computational resources

This paper describes a similarity measure for images which can be used in image-based topological localization and topological SLAM problems by autonomous robots with low computational resources. Instead of storing the images in the robot’s memory, we propose a compact signature to be extracted from the images. The signature is based on the calculation of the 2D Haar Wavelet Transform of the gray-level image and its size is only 170 bytes. We called this signature the DWT-signature. We exploit the frequency and space localization property of the wavelet transform to match the images grabbed by the perspective camera mounted on board the robot and the reference panoramic images built using an automatic image stitching procedure. The proposed signature allows, at the same time, memory saving and fast and efficient similarity calculation. For the topological SLAM problem we also present a simple implementation of a loop-closure detection based on the proposed signature.We report experiments showing the effectiveness of the proposed image similarity measure using two kinds of small robots: an AIBO ERS-7 robot of the RoboCup Araibo Team of the University of Tokyo and a Kondo KHR-1HV humanoid robot of the IAS-Lab of the University of Padua.     

WALRUS: a similarity retrieval algorithm for image databases

Approaches for content-based image querying typically extract a single signature from each image based on color, texture, or shape features. The images returned as the query result are then the ones whose signatures are closest to the signature of the query image. While efficient for simple images, such methods do not work well for complex scenes since they fail to retrieve images that match the query only partially, that is, only certain regions of the image match. This inefficiency leads to the discarding of images that may be semantically very similar to the query image since they may contain the same objects. The problem becomes even more apparent when we consider scaled or translated versions of the similar objects. We propose WALRUS (wavelet-based retrieval of user-specified scenes), a novel similarity retrieval algorithm that is robust to scaling and translation of objects within an image. WALRUS employs a novel similarity model in which each image is first decomposed into its regions and the similarity measure between a pair of images is then defined to be the fraction of the area of the two images covered by matching regions from the images. In order to extract regions for an image, WALRUS considers sliding windows of varying sizes and then clusters them based on the proximity of their signatures. An efficient dynamic programming algorithm is used to compute wavelet-based signatures for the sliding windows. Experimental results on real-life data sets corroborate the effectiveness of WALRUS'S similarity model.     

A high performance fingerprint liveness detection method based on quality related features

A new software-based liveness detection approach using a novel fingerprint parameterization based on quality related features is proposed. The system is tested on a highly challenging database comprising over 10,500 real and fake images acquired with five sensors of different technologies and covering a wide range of direct attack scenarios in terms of materials and procedures followed to generate the gummy fingers. The proposed solution proves to be robust to the multi-scenario dataset, and presents an overall rate of 90% correctly classified samples. Furthermore, the liveness detection method presented has the added advantage over previously studied techniques of needing just one image from a finger to decide whether it is real or fake. This last characteristic provides the method with very valuable features as it makes it less intrusive, more user friendly, faster and reduces its implementation costs.     

基于通道间相关性的图像重着色检测

图像重着色是一种新兴的图像编辑技术,通过篡改像素值达到改变图像颜色风格的目的。随着社交网络和图像编辑技术的快速发展,重着色图像已经严重阻碍了信息传达的真实性。然而,专门为重着色而设计的工作少之又少,现有的重着色检测方法在传统重着色场景下仍有很大提升空间,在应对手工重着色图像时效果不佳。为此,提出了一种基于通道间相关性的重着色图像检测方法,该方法适用于重着色任务中的传统重着色和手工重着色场景。基于相机成像和重着色图像生成方式之间存在显著差异这一现象,提出重着色操作或许会破坏自然图像的通道间相关性这一假设。通过数值分析说明,通道间相关性差异可作为区分重着色图像和自然图像的重要鉴别度量。基于上述先验知识,所提方法通过提取差分图像的一阶微分残差的通道共生矩阵,获得图像的通道间相关性特征集。此外,根据实际情况,假设了3种检测场景,包括训练-测试数据之间匹配、不匹配以及手工重着色场景。实验结果表明,所提方法能够准确识别重着色图像,在假设的3种场景下均优于现有方法,取得了较高的检测精度。除此之外,所提方法对训练数据量的依赖性较小,在训练数据有限的情况下,能实现相当精确的预测结果。     

A robust loop-closure method for visual SLAM in unstructured seafloor environments

This paper addresses the problem of visual simultaneous localization and mapping (SLAM) in an unstructured seabed environment that can be applied to an unmanned underwater vehicle equipped with a single monocular camera as the main measurement sensor. Monocular vision is regarded as an efficient sensing option in the context of SLAM, however it poses a variety of challenges when the relative motion is determined by matching a pair of images in the case of in-water visual SLAM. Among the various challenges, this research focuses on the problem of loop-closure which is one of the most important issues in SLAM. This study proposes a robust loop-closure algorithm in order to improve operational performance in terms of both navigation and mapping by efficiently reconstructing image matching constraints. To demonstrate and evaluate the effectiveness of the proposed loop-closure method, experimental datasets obtained in underwater environments are used, and the validity of the algorithm is confirmed by a series of comparative results.     

Face recognition in low resolution thermal images

This paper proposes an accurate, rotation invariant, and fast approach for detection of facial features from thermal images. The proposed approach combines both appearance and geometric information to detect the facial features. A texture based detector is performed using Haar features and AdaBoost algorithm. Then the relation between these facial features is modeled using a complex Gaussian distribution, which is invariant to rotation. Experiments show that our proposed approach outperforms existing algorithms for facial features detection in thermal images. The proposed approach’s performance is illustrated in a face recognition framework, which is based on extracting a local signature around facial features. Also, the paper presents a comparative study for different signature techniques with different facial image resolutions. The results of this comparative study suggest the minimum facial image resolution in thermal images, which can be used in face recognition. The study also gives a guideline for choosing a good signature, which leads to the best recognition rate.     

面向商用活体检测平台的鲁棒性评估

活体检测技术已经成为日常生活中的重要应用,手机刷脸解锁、刷脸支付、远程身份验证等场景都会用到这一技术。但如果攻击者利用虚假视频生成技术生成逼真的换脸视频来攻击上述场景的活体检测系统,将会对这些场景的安全性产生巨大的威胁。针对这个问题使用4种先进的Deepfake技术生成大量的换脸图片和视频作为测试样本,用这些样本来对百度、腾讯等商用活体检测平台的在线API接口进行测试。测试实验结果显示常用的各大商用活体检测平台对Deepfake图像的检测成功率普遍很低,并且对图像的质量较为敏感,对真实图像的误检率也很高。其主要原因可能是这些平台设计时针对的是打印照片攻击、屏幕二次翻拍攻击、硅胶面具攻击等传统的活体检测攻击方法,并未将先进的换脸检测技术集成到他们的活体检测算法中,这些平台因此不能够有效应对Deepfake攻击。因此,提出了一种集成活体检测方法Integranet,该方法由4种针对不同图像特征的检测算法集成所得,既能够有效检测出打印照片、屏幕二次翻拍等传统的攻击手段,也能够有效应对先进的Deepfake攻击。在测试数据集上验证Integranet的检测效果,结果显示Integranet检测...     

Off-line handwritten signature detection by analysis of evidence accumulation

One fundamental step in off-line handwritten signature verification is the detection of the signature position within the document image. This paper introduces an original approach for signature position detection. The method is based on an accumulative evidence technique, searching the region that maximizes some measure of correspondence with a given reference signature. This measure is based on the similarity of the slope marked out by each of the strokes in the signature. Experiments have shown that the method can be used on real documents, such as bank checks, where images have a high noise level due to background interferences (i.e. machine or handwritten texts, stamps, and lines). The proposed method is robust to variability in the size of the signatures and has the advantage of using only one reference signature per person.     

A fast face detection architecture for auto-focus in smart-phones and digital cameras一个快速的人脸检测架构用于智能手机和数码相机的自动对焦

Auto-focus is very important for capturing sharp human face centered images in digital and smart phone cameras. With the development of image sensor technology, these cameras support more and more highresolution images to be processed. Currently it is difficult to support fast auto-focus at low power consumption on high-resolution images. This work proposes an efficient architecture for an AdaBoost-based face-priority auto-focus. The architecture supports block-based integral image computation to improve the processing speed on high-resolution images; meanwhile, it is reconfigurable so that it enables the sub-window adaptive cascade classification, which greatly improves the processing speed and reduces power consumption. Experimental results show that 96% detection rate in average and 58 fps (frame per second) detection speed are achieved for the 1080p (1920×1080) images. Compared with the state-of-the-art work, the detection speed is greatly improved and power consumption is largely reduced.     

基于特征融合的可见光图像舰船检测新方法

该文以可见光图像舰船目标为研究对象,提出了用多谱图像和全色图像进行特征融合来检测舰船目标的方法。该方法首先利用多谱图像实现水域和陆地的分离,然后把分类结果映射到全色图像上从而实现在全色图像上区分水域和陆地;屏蔽陆地后用Otsu方法分别在多谱图像和全色图像上对目标进行分割,并提取目标特征,最后对目标特征进行融合来检测舰船目标。实验证明该方法有效可行。     

基于深度卷积神经网络的航空器检测与识别

针对军用机场大尺寸卫星图像中航空器检测识别的具体应用场景,建立了一套实时目标检测识别框架,将深度卷积神经网络应用到大尺寸图像中的航空器目标检测与识别任务中。首先,将目标检测的任务看成空间上独立的bounding-box的回归问题,用一个24层卷积神经网络模型来完成bounding-box的预测;然后,利用图像分类网络来完成目标切片的分类任务。大尺寸图像上的传统目标检测识别算法通常在时间效率上很难突破,而基于卷积神经网络的航空器目标检测识别算法充分利用了计算硬件的优势,大大缩短了任务耗时。在符合应用场景的自采数据集上进行测试,所提算法目标检测实时性达到平均每张5.765 s,在召回率65.1%的工作点上达到了79.2%的精确率,分类网络的实时性达到平均每张0.972 s,Top-1错误率为13%。所提框架在军用机场大尺寸卫星图像中航空器检测识别的具体应用问题上提出了新的解决思路,同时保证了实时性和算法精度。     

基于小波分层的多方向图像边缘检测

图像处理中, 边缘检测具有很重要的作用, 它可作为模式识别、图像分割及图像场景分析的基础. 传统的图像边缘算法具有算法简单, 方向适应性强的优势, 然而由于图像边缘具有多样性(方向的不一致性、边缘强弱的不相同等), 这些传统算法不能很好的体现出优越性. 本文结合目前先进的小波理论, 将图像进行小波变换, 得到具有单一性边缘的子图像, 再将传统边缘检测算子的方向性与这些子图像对应起来分别进行检测, 最后分别得到不同强度(层次)图像边缘, 并且这些边缘可以进行合成, 得到较好的图像边缘. 该算法操作简单, 具有很好的效果.     

基于目标检测的SAR图像变化检测方法

SAR图像变化检测有着广阔的应用前景,但目前的方法普遍以精确的配准为前提,使其适用范围受到限制。针对人造目标在SAR图像上的特点,提出了一种基于目标检测的SAR图像变化检测方法。通过图像中的人造目标之间相对位置关系的相似程度确定图像的变化情况,以此来降低对图像配准精度的要求。实验表明该方法在很宽的配准精度范围内都可获得较满意的结果。     

图像增强对显著性目标检测的影响研究

目的 雾霾、雨雪天气和水下等非理想环境因素会引起图像退化,导致出现低质图像,从而影响人类主观视觉感受及机器视觉应用任务的性能,因此,低质图像被利用之前进行图像增强成为惯常的预处理过程。然而,图像增强能否提高图像机器视觉应用任务的性能及影响程度等问题鲜有系统性研究。针对上述问题,本文以图像显著性目标检测这一机器视觉应用为例,研究图像增强对显著性目标检测性能的影响。方法 首先利用包括5种传统方法、6种深度学习方法等共11种典型图像增强方法对图像进行增强处理,然后利用8种典型的显著性目标检测方法对增强前后的图像分别进行显著性目标检测实验,并对比分析其结果。结果 实验表明,图像增强对低质图像显著性目标检测方法性能的促进作用不明显,某些增强方法甚至表现出负面影响,也存在同一增强方法对不同的显著性目标检测方法作用不同的现象。结论 图像增强对于显著性目标检测及其他的机器视觉应用的实际效果值得进一步研究,如何根据图像机器视觉应用的需求来选择和设计有效的增强方法需进一步探讨。     

一种基于数字签名的二值图像认证算法

为了实现对二值文本图像内容的全面保护,提出一种新的基于数字签名的二值图像认证算法。算法用数字签名实现对均匀块的认证,用归一化预处理实现对非均匀块的认证,通过两者联合认证来消除虚警。理论分析和实验结果表明,该算法仅需要附加极短的签名信息,就可以实现对二值文本图像内容的全面保护,在保证良好视觉效果的前提下,具有良好的窜改检测和窜改定位能力。     

结合数字水印和签名的二值图像内容验证

该文在笔者以往研究的基础上,针对二值图像的特点,提出了一种结合数字水印和数字签名的算法,将该算法应用于二值图像内容验证。算法先从二值图像中提取签名值,然后把签名值信息作为水印随机地嵌回原图像。在此过程中,用“可修改分值”对图像每个象素点的“可修改程度”进行了量化,然后根据“可修改分值”大小随机选择象素点进行水印嵌入。理论分析和实验结果都表明,算法具有很好的不可感知性和安全性(抗攻击性)。实验还表明,算法对于针对二值图像的微小窜改都具有足够的灵敏度。     

Edge Detection in Pipe Images Using Classification of Haar Wavelet Transforms

Automatic image interpretation for pipe inspection is a relatively recent area of research, which has great potential benefit. An important component of such systems is crack detection, or, more generally, edge or discontinuity detection. This paper describes a new approach to edge detection and applies it to pipe images. The method labels each pixel in an image as an edge pixel or a nonedge pixel by processing the Haar wavelet transform of the image in a window about the pixel using a support vector machine. As a pixel classifier, to within a moderate morphological tolerance, the detector has an accuracy of 99% on the images on which it has been tested and compares favorably with the commonly used Canny edge detector.     

自然场景下变形车牌检测模型DLPD-Net

目的 随着智能交通领域车牌应用需求的升级,以及车牌图像复杂性的提高,自然场景下的车牌识别面临挑战。为应对自然场景下车牌的不规则变形问题,充分考虑车牌的形状特征,提出了一种自然场景下的变形车牌检测模型DLPD-Net （distorted license plate detection network）。方法 该模型首次将免锚框目标检测方法应用于车牌检测任务中,不再使用锚框获取车牌候选区域,而是基于车牌热力值图与偏移值图来预测车牌中心;然后基于仿射变换寻找车牌角点位置,将变形车牌校正为接近于正面视角的平面矩形,从而实现在各种自然场景下变形车牌的检测。结果 一方面,基于数据集CD-HARD评估DLPD-Net检测算法的性能;另一方面,基于数据集AOLP （the application-oriented license plate database）和CD-HARD评估基于DLPD-Net的车牌识别系统的有效性。实验结果表明,DLPD-Net具有更好的变形车牌检测性能,能够提升车牌识别系统的识别准确率,在数据集CD-HARD上识别准确率为79.4%,高出其他方法4.4% 12.1%,平均处理时间为237 ms。在数据集AOLP上取得了96.6%的识别准确率,未使用扩充数据集的情况下识别准确率达到了94.9%,高出其他方法1.6% 25.2%,平均处理时间为185 ms。结论 本文提出的自然场景下的变形车牌检测模型DLPD-Net,能够实现在多种变形条件下的车牌检测,鲁棒性强,对遮挡、污垢和图像模糊等复杂自然环境下的车牌检测具有良好检测效果,同时,基于该检测模型的车牌识别系统在非受限的自然场景下具有更高的实用性。     

A survey of passive technology for digital image forensics

Over the past years, digital images have been widely used in the Internet and other applications. Whilst image processing techniques are developing at a rapid speed, tampering with digital images without leaving any obvious traces becomes easier and easier. This may give rise to some problems such as image authentication. A new passive technology for image forensics has evolved quickly during the last few years. Unlike the signature-based or watermark-based methods, the new technology does not need any signature generated or watermark embedded in advance. It assumes that different imaging devices or processing would introduce different inherent patterns into the output images. These underlying patterns are consistent in the original untampered images and would be altered after some kind of manipulations. Thus, they can be used as evidence for image source identification and alteration detection. In this paper, we will discuss this new forensics technology and give an overview of the prior literatures. Some concluding remarks are made about the state of the art and the challenges in this novel technology.