Satellite imagery is changing the way we understand and predict economic activity in the world. Advancements in satellite hardware and low-cost rocket launches have enabled near-real-time, high-resolution images covering the entire Earth. It is too labour-intensive, time-consuming and expensive for human annotators to analyse petabytes of satellite imagery manually. Current computer vision research

Facial attribute editing has gained increasing attention recently. Previous methods tackle this challenge by incorporating encoder–decoder and generative adversarial networks. However, the bottleneck layer in encoder–decoder of these methods often leads to blurry and low-quality editing results. And skip connections are used between deep and shallow layers to improve image quality but suffer from a

Automatic optical inspection is a vital part of the production process for solder joints appearance inspection in surface mounted technology assembling lines. However, IC solder joint inspection is a challenging problem because IC solder joints have extremely small sizes and no distinct appearance differences between qualified and unqualified ones. In this paper, we propose an IC solder joint inspection

Estimating the accuracy of subpixel image registration algorithms in practical experiments is important, yet challenging in many applications. The accuracy of these algorithms in camera-based systems is influenced not only by image content, but also camera noise. In this study, five experiments were designed to quantify the effect of changing camera settings on the image noise level, and consequently

A correction to this paper has been published: https://doi.org/10.1007/s00138-021-01222-5

Training a deep learning model for identifying dangerous vehicles requires a large amount of labeled accident data. However, it is difficult to collect a sufficient amount of accident data in the real world. To address this challenge, we introduce a driving-simulator-based data generator that can arbitrarily produce a wide variety of accident scenarios. Furthermore, in order to reduce the gap between

It is often difficult to obtain the high-precision inner cavity contour size and 3D model of parts and components in reverse engineering. This paper proposes a method that uses a sequence of section images of a part to reconstruct their 3D models. This method cuts the part layer by layer to obtain the sectional images and extracts the 3D information of the sectional image contours to generate point

Capsule network (CapsNet) is a novel concept demonstrating the importance of learning spatial hierarchical relationship between features for the effective recognition of images. However, the baseline capsule network is not suitable for the recognition of complex images leading to its poor performance on such images. This limitation can partially be attributed to the inability of CapsNets to extract

In this work, we face the problem of estimating the relative position and orientation of a camera and an object, when they are both equipped with inertial measurement units (IMUs), and the object exhibits a set of n landmark points with known coordinates (the so-called Pose estimation or PnP Problem). We present two algorithms that, fusing the information provided by the camera and the IMUs, solve

Despite the considerable advances that are emerged in correlation filter-based tracking, in fact, they may achieve excellent performance in robustness, speed, and accuracy; they still fail when dealing with large-scale alteration and show the inability to handle long-term tracking in complex scenarios, where the object undergoes partial occlusion, out-of-view, and deformation. In this paper, we propose

A multi-spectral imaging technique for the swift fusion of red–green–blue (RGB) and near infrared (NIR) image pairs with a deep learning based resolution enhancement technique is proposed, mpirically investigated and compared to some state-of-the-art techniques in the current work. The results of the proposed multi-spectral image fusion demonstrate good chrominance preservation, improved sharpness

Effective bounding box regression is essential for running any real-time object detection algorithm with acceptable accuracy. The currently available loss functions have issues like high computations, and sometimes they suffer from a subtle problem of plateau for non-overlapping bounding boxes, as the resultant bounding boxes are found to be far from the ground truth. In the present investigation,

Image inpainting and denoising are two important preprocessing steps widely used in image and visual analysis. In this paper, by the maximum a posterior estimation, we present a new framework to remove multiplicative noise and artifacts simultaneously, when the locations of the artifacts/damaged pixels are unknown. By taking into account the statistical distribution of multiplicative noise as Gamma

Person re-identification is an important video analysis problem that aims to track people over non-overlapping views in a large camera network. The purpose is to find the same person from disjoint camera views at different times and locations. Most of the existing re-identification approaches assume pre-selected people bounding box captured by two non-overlapping cameras to perform the re-identification

High resolution electroluminescence (EL) images captured in the infrared spectrum allow to visually and non-destructively inspect the quality of photovoltaic (PV) modules. Currently, however, such a visual inspection requires trained experts to discern different kinds of defects, which is time-consuming and expensive. Automated segmentation of cells is therefore a key step in automating the visual

For the past several years, a variety of algorithms have been focused on how to exploit two-dimensional scanline optimization to augment one-dimensional ones for semi-global matching. Different from the former contributions, an ensemble learning with advanced fast image filtering features for semi-global matching is proposed in this paper. Firstly, fewer categories of features (confidence measures)

We propose a novel approach that addresses face aging as an unsupervised image-to-image translation problem. The proposed approach achieves age progression (i.e., future looks) and regression (i.e., previous looks) of face images that belong to a specific age class by translating them to other (subsequent or precedent) age classes. It learns pairwise translations between all age classes. Two variants

Bundle adjustment is a least squares method-based algorithm for minimizing the global reprojection error and has provided an effective solution for structure from motion (SfM). The Levenberg–Marquardt algorithm provides a feasible and convenient way for bundle adjustment and creates a system of linear equations which are normal equations. For the special sparsity structure of an augmented Hessian matrix

Recently, surveillance systems have been widely used to analyze video recordings captured by surveillance cameras and to detect abnormal or irregular events in real-world scenes. In this study, we present a novel system that detects abnormal events. Unlike conventional methods, we consider abnormal event detection as variation matching problems. In approaching this problem, we transform from a single

Deep-fried carbohydrate-rich foods items such as potato chips and French fries are one of the most popular snack foods consumed across the globe. In the production of these carbohydrate-rich foods items, a compound known as acrylamide is formed which is carcinogen and mutagen as well. The conventional chemical-based methods for detection of the presence of acrylamide in the deep-fried carbohydrate-rich

The present paper aims to study image processing algorithms to accelerate and facilitate the evaluation of the harvest condition in tomato farms. In order to achieve this, two different deep learning models are trained and combined with counting methods to produce a harvest monitoring system for embedded applications using an Intel® MovidiusTM and an affordable RGB camera. The first model detects the

The tunnel boring machine (TBM) is a key equipment for excavating long-range tunnels. It is a complex system and hard to be controlled well in practice. In this paper, we propose the rock segmentation visual system to assist TBM driving. Through the system, online size distribution of excavated rocks is automatically analysed and sent back to TBM driver, from which many statistical information can

Over recent years, many approaches have been proposed for the denoising or semantic segmentation of X-ray computed tomography (CT) scans. In most cases, high-quality CT reconstructions are used; however, such reconstructions are not always available. When the X-ray exposure time has to be limited, undersampled tomograms (in terms of their component projections) are attained. This low number of projections

In this paper, we propose a novel video reconstruction methodology built based on a generalization of alternating direction method of multipliers (ADMM) named Plug-and-Play. The motivation of the proposed technique is the improvement in visual quality performance of the video frames and decreasing the reconstruction error in comparison with the former video reconstruction methods. The proposed algorithm

Image co-segmentation is important for its advantage of alleviating the ill-pose nature of image segmentation through exploring the correlation between related images. Many automatic image co-segmentation algorithms have been developed in the last decade, which are investigated comprehensively in this paper. We firstly analyze visual/semantic cues for guiding image co-segmentation, including object

In ship classification, selecting distinctive features and designing a proper classifier are two key points of the process. As a lack of most of the studies, these two essential points are considered separately. In this study, our proposal includes joint feature extraction, selection, and classifier design framework to build a novel deep cascade network for ship classification. We propose a transfer

In the automotive industry, light-alloy aluminum castings are an important element for determining roadworthiness. X-ray testing with computer vision is used during automated inspections of aluminum castings to identify defects inside of the test object that are not visible to the naked eye. In this article, we evaluate eight state-of-the-art deep object detection methods (based on YOLO, RetinaNet

Maritime vessel re-identification (re-ID) is a computer vision task of vessel identity matching across disjoint camera views. Prominent applications of vessel re-ID exist in the fields of surveillance and maritime traffic flow analysis. However, the field suffers from the absence of a large-scale dataset that enables training of deep learning models. In this study, we present a new dataset that includes

We propose a new hierarchical structure from motion system for close loop sequences. Our system includes a novel approach for clustering cameras into multiple sets, whose camera poses are initially reconstructed separately and later globally registered w.r.t a single coordinate frame. Each of the multiple sets is robustly reconstructed by a novel guarded least median square protocol. Our method is

Studies on contextual abnormality detection and localization for images and videos are presented in this work. The task of detecting abnormalities becomes challenging while considering the context in the scene. Some object which is normal in one scenario may be considered as abnormal in another. We present conceptually simple, flexible and a general framework, by incorporating instance segmentation

A correction to this paper has been published: https://doi.org/10.1007/s00138-021-01200-x

Deep neural networks are vulnerable to adversarial examples, i.e., carefully perturbed inputs designed to mislead the network at inference time. Recently, adversarial patch, with perturbations confined to a small and localized patch, emerged for its easy accessibility in real-world attack. However, existing attack strategies require training data on which the deep neural networks were trained, which

In this paper, we propose an AI (Artificial Intelligence) solution for solder ball HIP (Head-In-Pillow) defect inspection. The HIP defect will affect the conductivity of the solder balls leading to intermittent failures. Due to the variable location and shape of the HIP defect, traditional machine vision algorithms cannot solve the problem completely. In recent years, Convolutional Neural Network (CNN)

Depth degradation caused by the conditions and environment of depth sensor hardware restricts its application potential, and this limitation cannot be avoided simply by improving the design of sensor. To overcome this limitation, we propose a multi-resolution depth image restoration method. Firstly, the sub-images of depth image and color image at different scales are obtained by multi-resolution analysis

Image will be seriously distorted when camera captures the scene through water surface. This will cause the captured image to lose its readability. Although several previous methods have been proposed to resolve this problem, the methods still need improvement. In this paper, we propose a reconstruction method to restore the distorted images based on a twice registration algorithm. Several calculation

This work introduces a new approach to localize anomalies in surveillance video. The main novelty is the idea of using a Siamese convolutional neural network to learn a metric between a pair of video patches (spatiotemporal regions of video). The learned metric, which is not specific to the target video, is used to measure the perceptual distance between each video patch in the testing video and the

Single-image dehazing is an extensively studied field and an ill-posed problem faced by vision-based systems in an outdoor environment. This paper proposes a dark channel network to estimate the transmission map of an input hazy scene for single-image dehazing. The architecture constitutes two major components—feature extraction layer and convolutional neural network layer. The former extracts the

The objective of this paper is to argue that data mattered more than network in terms of color constancy. Computational color constancy is a linear operation device-dependent, which is part of the camera imaging pipeline. We extend the dataset based on this pipeline and prove that the scene illumination can be predicted using a very simple network as long as the dataset is large enough and evenly distributed

The detection of defects in solar cells based on machine vision has become the main direction of current development, but the graphical feature extraction of micro-cracks, especially cracks with complex shapes, still faces formidable challenges due to the difficulties associated with the complex background, non-uniform texture, and poor contrast between crack defects and background. In this paper,

To design a hybrid deep learning system (hDL-system) for discriminating low-grade from high-grade colorectal cancer (CRC) lesions, using immunohistochemically stained biopsy specimens for AIB1 expression. AIB1 has oncogenic function in tumour genesis, and it is an important prognostic factor regarding various types of cancers, including CRC. Clinical material consisted of biopsy specimens of sixty-seven

Siamese trackers demonstrated high performance in object tracking due to their balance between accuracy and speed. Unlike classification-based CNNs, deep similarity networks are specifically designed to address the image similarity problem and thus are inherently more appropriate for the tracking task. However, Siamese trackers mainly use the last convolutional layers for similarity analysis and target

Detecting a panic behavior within a human crowd is of a high importance as it allows to prevent disasters. Online analysis of video streams, real-time processing and accurate detection are required to ensure effective surveillance of the crowded places. However, these requirements are not simultaneously fulfilled by the existing techniques. Rapid advances in artificial intelligence are investing the

Finger knuckle print (FKP) as a physiological trait with a small image dimension, also a highly distinctive pattern, can be used as a reliable biometric identifier. In this paper, a new effective biometric authentication system using FKP texture based on relaxed local ternary pattern (RLTP) is presented. To further improve performance, cascading, overlapped patching and uniform rotation invariant pattern

Given single instance (or template image) per product, our objective is to detect merchandise displayed in the images of racks available in a supermarket. Our end-to-end solution consists of three consecutive modules: exemplar-driven region proposal, classification followed by non-maximal suppression of the region proposals. The two-stage exemplar-driven region proposal works with the example or template

To tackle problems arising from unexpected camera motions in unmanned aerial vehicles (UAVs), we propose a three-mode ensemble tracker where each mode specializes in distinctive situations. The proposed ensemble tracker is composed of appearance-based tracking mode, homography-based tracking mode, and momentum-based tracking mode. The appearance-based tracking mode tracks a moving object well when

Pairwise frame registration with sparse geometric local features on real-world depth images is not particularly robust due to the low resolution and incomplete nature of the 3D scan data. Moreover, there might be many regions with similar geometric information. In this paper, we present 3DTDesc, a data-driven descriptor which closely combines both 2D texture and 3D geometric information for frame registration

Understanding the complex semantic structure of scene images requires mapping the image from pixel space to high-level semantic space. In semantic space, a scene image is represented by the posterior probabilities of concepts (e.g., ‘car,’ ‘chair,’ ‘window,’ etc.) present in it and such representation is known as semantic multinomial (SMN) representation. SMN generation requires a concept annotated

Powder metallurgy gears are often accompanied by broken teeth, abrasion, scratches and crack defects. In order to eliminate the defective gears in gear production and improve the yield of gears, this paper presents an improved GA–PSO algorithm, called the SHGA–PSO algorithm. Firstly, the gear images were preprocessed by bilateral filtering, and the images were segmented by the Sobel operator. Then

In this paper, we present a novel machine learning-based image ranking approach using Convolutional Neural Networks (CNN). Our proposed method relies on a similarity metric learning algorithm operating on lists of image examples and a loss function taking into account the ranking in these lists with respect to different query images. This comprises two major contributions: (1) Rank lists instead of

Multi-exposure fusion is a technique used to generate a high-dynamic-range image without calculating the camera response function and without compressing its ranges with the tone mapping process. There are many schemes for fusing multi-exposure images. One of the famous schemes is the pyramid-based blending, which fuses multi-exposure images together based on the concept of multi-resolution blending

Small inter-class and large intra-class variations are the key challenges in fine-grained visual classification. Objects from different classes share visually similar structures, and objects in the same class can have different poses and viewpoints. Therefore, the proper extraction of discriminative local features (e.g., bird’s beak or car’s headlight) is crucial. Most of the recent successes on this

This paper introduces a novel method to super-resolve multi-spectral images captured by modern real-time single-shot mosaic image sensors, also known as multi-spectral cameras. Our contribution is twofold. Firstly, we super-resolve multi-spectral images from mosaic images rather than image cubes, which helps to take into account the spatial offset of each wavelength. Secondly, we introduce an external

Pedestrian detection and re-identification technology is a research hotspot in the field of computer vision. This technology currently has issues such as insufficient pedestrian expression ability, occlusion, diverse pedestrian attitude, and difficulty of small-scale pedestrian detection. In this paper, we proposed an end-to-end pedestrian detection and re-identification model in real scenes, which

Salient object detection is a hot spot of current computer vision. The emergence of the convolutional neural network (CNN) greatly improves the existing detection methods. In this paper, we present 3MNet, which is based on the CNN, to make the utmost of various features of the image and utilize the contour detection task of the salient object to explicitly model the features of multi-level structures

This study aims to utilize a modified you only look once (YOLO) network to address the detection and classification of spilled loads on freeways. YOLO architecture was augmented in two ways. Firstly, a kernel size of 1 × 1 for the conv layers was used. Secondly, the use of connections between the convolution layers was proposed. For training the network, a synthetic dataset was constructed where ImageNet

Crowd density classification has been a challenging task in the field of computer vision, which has various applications in public and commercial domains. Many researches on the classification and recognition method of the crowd density have been introduced in the past, while there still exists the problems of inaccuracy, poor robustness and inefficiency. An adaptive crowd density classification method

Automated three-dimensional (3D) object reconstruction is the task of building a geometric representation of a physical object by means of sensing its surface. Even though new single-view reconstruction techniques can predict the surface, they lead to incomplete models, specially, for non-commons objects such as antique objects or art sculptures. Therefore, to achieve the task’s goals, it is essential

Most domain adaptation methods consider the problem of transferring knowledge to the target domain from a single-source dataset. However, in practical applications, we typically have access to multiple sources. In this paper we propose the first approach for multi-source domain adaptation (MSDA) based on generative adversarial networks. Our method is inspired by the observation that the appearance

The Gaussian kernel function is widely used to encode the nonlinear correlations of the face images. However, some issues greatly limit its superiority, for example, it is sensitive to the parameter setting because of its definition based on the exponential operation, on the other hand, the Gaussian kernel needs costly computational time. Besides, the hidden information such as the distance information

Coded targets are used as reference targets with a known location during camera calibration, for robust searching of corresponding features between images during various applications of machine vision like object tracking, robot navigation or 3D measurement. In this paper, a target system for industrial photogrammetric applications is outlined. The methods which have been chosen emphasize maximum robustness