Search

Copyright of Journal of Test & Measurement Technology is the property of Publishing Center of North University of China and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

The difficulties in communication and hearing are an important concern for deaf–dumb people, which stop access to their essential and basic needs. Many findings have been made to address sign languages even though this challenging problem is not still solved. Many methods aimed to propose vision-based classifiers through identical pattern investigation tasks by obtaining the difficult handcraft feature descriptions of gestures from the gathered images. However, the efficacy of all those models is less for performing with a huge signbook captured from uncontrolled and complex background conditions. So, an effective Indian Sign Language (ISL) classification method is developed by an advanced deep learning approach. At first, the hand gesture images are obtained from the data source. Only the image of the hand, even from a complicated background, is extracted from the obtained image. The features are extracted using the Scale-Invariant Feature Transform (SIFT) method and Multiscale Vision Transformer (MVT). Then, the extracted features are fed to the Hybrid Convolution-based EfficientNet (HCEN) model. The hyper-parameters in the developed HCEN model are tuned using the implemented Adaptive Political Optimizer (APO) algorithm. The recognized hand signs are obtained from the suggested HCEN model. Various experiments are conducted to determine the performance of the suggested deep learning-based hand gesture recognition model. [ABSTRACT FROM AUTHOR]

The thickness of the bonding interface changes after the structure has been subjected to tensile load, thereby affecting the strength of the bonding structure. It is necessary to monitor the evolution of the bonding interface thickness under tensile load. Through terahertz time-domain spectroscopy imaging and scale-invariant feature transform image processing, the thickness variation of three interfaces of 150 silicon-based adhesive structures under tensile load were analyzed. It was determined that variation at the edge was larger than that in the middle area. The research results also provide data support for the bonding structure health monitoring systems. [ABSTRACT FROM AUTHOR]

Robust watermarking plays a key role in copyright protection and information leakage tracking, network conference recording, etc. However, cropping attack is a strong geometric attack, which can destroy the watermarked information of audio signal. For this problem, this paper proposes a robust audio watermarking algorithm by combining chaotic system, scale invariant feature transform (SIFT) feature and discrete wavelet transform (DWT) to resist cropping attack. After applying DWT to the original signal, we use the obtained DWT low-frequency coefficients to construct a square matrix and extract the SIFT features with scale invariance from this matrix. The watermark is encrypted by Tent map, and then it is embedded into the low-frequency components according to the sift features location information. Watermark extraction is the inverse process of embedding. Compared with the state of the art audio watermarking algorithms, the proposed algorithm has better performance in terms of robustness and payload capacity while ensuring good imperceptibility. [ABSTRACT FROM AUTHOR]

Individual recognition of animals is an important aspect of ecological sciences. Photograph‐based individual recognition options are of particular importance since these represent a non‐invasive method to distinguish and identify individual animals. Recent developments and improvements in computer‐based approaches make possible a faster semi‐automated evaluation of large image databases than was previously possible. We tested the Scale Invariant Feature Transform (SIFT) algorithm, which extracts distinctive invariant features of images robust to illumination, rotation or scaling of images. We applied this algorithm to a dataset of 800 tail pattern images from 100 individual Eurasian beavers (Castor fiber) collected as part of the Norwegian Beaver Project (NBP). Images were taken using a single‐lens reflex camera and the pattern of scales on the tail, similar to a human fingerprint, was extracted using freely accessible image processing programs. The focus for individual recognition was not on the shape or the scarring of the tail, but purely on the individual scale pattern on the upper (dorsal) surface of the tail. The images were taken from two different heights above ground, and the largest possible area of the tail was extracted. The available data set was split in a ratio of 80% for training and 20% for testing. Overall, our study achieved an accuracy of 95.7%. We show that it is possible to distinguish individual beavers from their tail scale pattern images using the SIFT algorithm. [ABSTRACT FROM AUTHOR]

Purpose : Induction Motor (IM) is used in various industrial application such as drilling, rolling, paper mills, lathe machine and milling. The faults in IM are avoided through continuous monitoring and control, during running conditions. This research paper presents an innovative method for detecting motor faults such as unbalanced voltage, phase reversal, frequent starting, abnormal speed, low voltage, phase loss, turn-to-turn short circuit, shaft bend and predicting Rotor Slot Size Variation (RSSV) during motor faults. Methods: The RSSV and rotor faults of the IM are due to magnetic and Thermal Stress (TS). Magnetic and thermal stresses are measured through GMR, and temperature sensor. The vibration sensor measures abnormal motor vibration during rotor faults. In this proposed method, Least Square Support Vector Machine (LS-SVM) and Symmetrized Dot Pattern (SDP) are used to match the sensor signals of normal and abnormal motors and identify motor faults. Results: The signals collected from IM are changed into 2D images using the concept of SDP, which shows the difference among various induced faults through dot patterns of signal data points. Energy band values are obtained from the GMR sensor signal using Scale Invariant Feature Transform (SIFT) and Microscopic Camera Image Values (MCIV) are provided to the Multiple Linear Regression (MLR) for predicting the RSSV. Conclusion : The prediction accuracy of rotor slot size through the proposed technique is 94.8%. From the experimental results, if the RSSV is more than 2% of the standard size, leads to heavy damage to the IM. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Education & Science is the property of Republic of Iraq Ministry of Higher Education & Scientific Research (MOHESR) and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Primary spontaneous pneumothorax (PSP) primarily affects slim and tall young males. Exploring the etiological link between chest wall structural characteristics and PSP is crucial for advancing treatment methods. In this case-control study, chest computed tomography (CT) images from patients undergoing thoracic surgery, with or without PSP, were analyzed using Artificial Intelligence. Convolutional Neural Network (CNN) model of EfficientNetB3 and InceptionV3 were used with transfer learning on the Imagenet to compare the images of both groups. A heatmap was created on the chest CT scans to enhance interoperability, and the scale-invariant feature transform (SIFT) was adopted to further compare the image level. A total of 2,312 CT images of 26 non-PSP patients and 1,122 CT images of 26 PSP patients were selected. Chest-wall apex pit (CAP) was found in 25 PSP and three non-PSP patients (p

Abstract Individual recognition of animals is an important aspect of ecological sciences. Photograph‐based individual recognition options are of particular importance since these represent a non‐invasive method to distinguish and identify individual animals. Recent developments and improvements in computer‐based approaches make possible a faster semi‐automated evaluation of large image databases than was previously possible. We tested the Scale Invariant Feature Transform (SIFT) algorithm, which extracts distinctive invariant features of images robust to illumination, rotation or scaling of images. We applied this algorithm to a dataset of 800 tail pattern images from 100 individual Eurasian beavers (Castor fiber) collected as part of the Norwegian Beaver Project (NBP). Images were taken using a single‐lens reflex camera and the pattern of scales on the tail, similar to a human fingerprint, was extracted using freely accessible image processing programs. The focus for individual recognition was not on the shape or the scarring of the tail, but purely on the individual scale pattern on the upper (dorsal) surface of the tail. The images were taken from two different heights above ground, and the largest possible area of the tail was extracted. The available data set was split in a ratio of 80% for training and 20% for testing. Overall, our study achieved an accuracy of 95.7%. We show that it is possible to distinguish individual beavers from their tail scale pattern images using the SIFT algorithm.

Copyright of Tunnel Construction / Suidao Jianshe (Zhong-Yingwen Ban) is the property of Tunnel Construction Editorial Office and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

This paper presents a new technique by support vector machines after extracting the prominent iris features using discrete wavelet transformations, to achieve an optimal classification of energy system disturbances. A framework for iris recognition and protection of the recognition system from fake iris scenes was proposed. The scale-invariant feature transform is set as an algorithm to extract local features (key points) from iris images and their classification method. To elicit the prominent iris features, the test image is first pre-processed. This will facilitate confining and segmenting the region of interest hopefully, reducing the blurring and artifacts, especially those associated with the edges. The textural features can be exploited to partition irises into regions of interest in addition to providing necessary information in the spatial distribution of intensity levels in an iris neighborhood. Next, the detection efficiency of the proposed method is achieved through extracting iris gradients and edges in complex areas and in different orientations. Further, the iris diagonal edges were easily detected after calculating the variance of different blocks in an iris and the additive noise variance in a textured image. The vertical, horizontal, and diagonal iris image gradients with different directions were successfully extracted. These gradients were extracted after adjusting the threshold amplitude obtained from the histograms of these gradients. The average calculations of MAVs, peak signal-to-noise ratios (PSNRs), and mean square errors (MSEs) within the orientation angles (?45°, +45° and 90°) for both vertical and horizontal iris gradients had occurred within the rates of 1.9455—3.1266, 36.388—39.863 dB and 0.0001—0.0026, respectively.

Facial Emotion Recognition (FER) is the approach of detecting emotions of humans from facial expressions. Emotions are detected automatically by the human brain and hence, many computer aided techniques are implemented for recognizing emotions. Recognition of natural emotions is an exciting area with a broad range of real time applications like automated tutoring models, smart environments computer–human interaction, driver warning systems, and video and image retrieval. The major goal of this paper is to implement an intelligent model for FER. The proposed model involves few steps: (a) Face extraction (b) Image filtering, (c) Extraction of facial components, (d) Descriptor selection, and (d) Classification. Initially, the face is extracted from the input image by the Viola–Jones method, which is generally employed for the detection of object. Further, the noise of the image is filtered by the Gabor Filtering. Further, the facial components are extracted as features via the Affine-Scale-Invariant Feature Transform (ASIFT), which is the modified version of SIFT method. Since the length of the descriptors generated from ASIFT is high; the number of descriptors is reduced by the optimal descriptor selection approach with the aid of hybrid meta-heuristic algorithm termed as MV-WOA. The extracted descriptors are subjected to Neural Network (NN). As a modification to the existing machine learning algorithms, the number of hidden neurons in the NN is optimized by the proposed WOA+MVO algorithm. By conducting experiments, the results demonstrate that the developed system can perform better the recent traditional approaches for emotion recognition for classifying seven emotions like normal, smile, sad, surprise, anger, fear, and disgust. [ABSTRACT FROM AUTHOR]

The commonly used artificial visual interpretation and existing object-oriented computer automatic recognition methods have some disadvantages, such as low efficiency and insufficient accuracy in recognizing earthquake-induced building damage in unmanned aerial vehicle (UAV)-based images. In this paper, we report the latest progress in research on machine learning algorithms in artificial intelligence, then propose a new method of recognizing earthquakeinduced building damage. Using the bag of words (BoW) model, scale-invariant feature transformation (SIFT) characteristics were clustered to build an eigenvector tag library with K clustering centers as visual words. After images were expressed by visual words as eigenvectors with unified dimensions, a histogram intersection kernel (HIK) was then employed to construct the histogram intersection kernel support vector machine (HIK-SVM) to classify images and recognize earthquake-induced building damage. Building damage due to the magnitude 6.0 earthquake that occurred in Luxian, Sichuan Province on September 16, 2021 was analyzed as an example. When the proposed method was applied to recognize earthquake damage using UAV-based images, the average recognition accuracy reached 91.7%. The experimental results verified the feasibility and validity of the proposed method. [ABSTRACT FROM AUTHOR]

Agriculture has become the main sources of the income for many developed countries. The productivity in agriculture can be affected by various diseases present in the plant due to climatic conditions. The key step to improve the productivity of crops are to detect the disease at the preliminary stage. Automation becomes the best solution for this because it is more difficult to observe the disorders in plants parts. For that an image of affected plant leaf is acquired and segments the affected portion and to recognize the disease by using image processing and computer vision and machine learning techniques. The extracted features from the segmented portion are descripted using Global and Local Visual descriptors. Finally, we use the classifier to recognize the disease. Extracting a meaningful feature from an image is a central problem for a variety of computer vision problems like recognition, image retrieval, and classification. In this research, visual feature descriptor that best describe an image with respect to its visual property is explored. It is specifically focusing on recognizing tasks. The experimental results have proved that the combination of visual descriptors with various classifiers such as SVM and Ensemble Classifier produces high quality outcomes when compared to individual descriptors.

Automated accurate categorization of brain magnetic resonance images (MRI) is very important for disease diagnosis and treatment. In this paper, a new methodology for the detection of abnormality in brain MRI is suggested. The scale-invariant feature transform is first employed to extract features of MRI. Principal component analysis is applied to the extracted features and a minimal set of more essential features is obtained. Lastly, the obtained feature set is categorized as healthy or unhealthy using support vector machine (SVM)-based classification. The parameters of SVM, i.e. C and s are optimized using Gray Wolf Optimization. However external noise and patient/organ movement degrade the quality of MRI, which in turn affect the classification accuracy. Therefore, a hybrid of Savitzky-Golay smoothing filter and speckle reducing anisotropic diffusion filter is used for preprocessing of the source image, which efficiently reduces the noise while preserving edges of the image. It is revealed from the results that proposed technique provides a classification accuracy of 99.61%. Thus the suggested technique may effectively diagnose diseases using MRI. [ABSTRACT FROM AUTHOR]

The study was conducted on several algorithms for determining similar images. Scale-invariant feature transform algorithm was taken as the main algorithm relative to which the change in performance indicators was calculated. The new algorithm is obtained by upgrading the main algorithm with the idea of comparing images using perceptual hashing. The analysis of the results was carried out and conclusions were drawn about the operation of the new algorithm, as well as limitations on its application. [ABSTRACT FROM AUTHOR]

The aim and objectives of the study are as follows: (i) to implement automated patch-based classification of hand X-ray images using modified pre-trained convolutional neural network (CNN) models; (ii) to develop a customized CNN model for automated feature extraction and classification of hand X-ray images and to compare the performance of customized CNN models with non-linear and linear kernels; (iii) to construct the hand crafted feature fusion (SIFT+ Customized CNN features) and categorize the normal and RA using Machine Learning classifiers. The model was trained on 75 images (10,000 patches) of hand radiographs and tested using 25 images (500 patches) that were not included in the training set. The accuracy of the modified pre-trained model GoogLeNet was 89% and the proposed custom model three achieved an accuracy of 95%. The sensitivity and specificity of GoogLeNet were 84% and 90% respectively. The custom model three attained the sensitivity and specificity as 95% and 94% respectively. Furthermore, when compared to the features extracted (SIFT + CNN) from the customized models, the custom3 model outperformed well for the classification of RA compared to ML classifiers. Thus a custom CNN-based computer-aided diagnostic tool can be used as an effective method for the detection of RA. [ABSTRACT FROM AUTHOR]

This paper presents a new technique by support vector machines after extracting the prominent iris features using discrete wavelet transformations, to achieve an optimal classification of energy system disturbances. A framework for iris recognition and protection of the recognition system from fake iris scenes was proposed. The scale-invariant feature transform is set as an algorithm to extract local features (key points) from iris images and their classification method. To elicit the prominent iris features, the test image is first pre-processed. This will facilitate confining and segmenting the region of interest hopefully, reducing the blurring and artifacts, especially those associated with the edges. The textural features can be exploited to partition irises into regions of interest in addition to providing necessary information in the spatial distribution of intensity levels in an iris neighborhood. Next, the detection efficiency of the proposed method is achieved through extracting iris gradients and edges in complex areas and in different orientations. Further, the iris diagonal edges were easily detected after calculating the variance of different blocks in an iris and the additive noise variance in a textured image. The vertical, horizontal, and diagonal iris image gradients with different directions were successfully extracted. These gradients were extracted after adjusting the threshold amplitude obtained from the histograms of these gradients. The average calculations of MAVs, peak signal-tonoise ratios (PSNRs), and mean square errors (MSEs) within the orientation angles (-45°, +45° and 90°) for both vertical and horizontal iris gradients had occurred within the rates of 1.9455--3.1266, 36.388--39.863 dB and 0.0001--0.0026, respectively. [ABSTRACT FROM AUTHOR]

When using drone-based aerial images for panoramic image generation, the unstableness of the shooting angle often deteriorates the quality of the resulting image. To prevent these polluting effects from affecting the stitching process, this study proposes deep learning-based outlier rejection schemes that apply the architecture of the generative adversarial network (GAN) to reduce the falsely estimated hypothesis relating to a transform produced by a given baseline method, such as the random sample consensus method (RANSAC). To organize the training dataset, we obtain rigid transforms to resample the images via the operation of RANSAC for the correspondences produced by the scale-invariant feature transform descriptors. In the proposed method, the discriminator of GAN makes a pre-judgment of whether the estimated target hypothesis sample produced by RANSAC is true or false, and it recalls the generator to confirm the authenticity of the discriminator's inference by comparing the differences between the generated samples and the target sample. We have tested the proposed method for drone-based aerial images and some miscellaneous images. The proposed method has been shown to have relatively stable and good performances even in receiver-operated tough conditions. [ABSTRACT FROM AUTHOR]

Content-Based Image Retrieval (CBIR) is an approach of retrieving similar images from a large image database. Recently CBIR poses new challenges in semantic categorization of the images. Different feature extraction technique have been proposed to overcome the semantic breach problems, however these methods suffer from several shortcomings. This paper contributes an image retrieval system to extract the local features based on the fusion of scale-invariant feature transform (SIFT) and KAZE. The strength of local feature descriptor SIFT complements global feature descriptor KAZE. SIFT concentrates on the complete region of an image using high fine points of features and KAZE ponders on details of a boundary. The fusion of local feature descriptor and global feature descriptor boost the retrieval of images having diverse semantic classification and also helps in achieving the better results in large scale retrieval. To enhance the scalability of image retrieval bag of visual words (BoVW) is mainly used. The fusion of local and global feature representations are selected for image retrieval for the reason that SIFT effectively captures shape and texture and robust towards the change in scale and rotation, while KAZE have strong response towards boundary and changes in illumination. Experiments conducted on two image collections, namely, Caltech-256 and Corel 10k demonstrate the proposed scheme appreciably enhanced the performance of the CBIR compared to state-of-the-art image retrieval techniques. [ABSTRACT FROM AUTHOR]

Agriculture has become the main sources of the income for many developed countries. The productivity in agriculture can be affected by various diseases present in the plant due to climatic conditions. The key step to improve the productivity of crops are to detect the disease at the preliminary stage. Automation becomes the best solution for this because it is more difficult to observe the disorders in plants parts. For that an image of affected plant leaf is acquired and segments the affected portion and to recognize the disease by using image processing and computer vision and machine learning techniques. The extracted features from the segmented portion are descripted using Global and Local Visual descriptors. Finally, we use the classifier to recognize the disease. Extracting a meaningful feature from an image is a central problem for a variety of computer vision problems like recognition, image retrieval, and classification. In this research, visual feature descriptor that best describe an image with respect to its visual property is explored. It is specifically focusing on recognizing tasks. The experimental results have proved that the combination of visual descriptors with various classifiers such as SVM and Ensemble Classifier produces high quality outcomes when compared to individual descriptors. [ABSTRACT FROM AUTHOR]

In recent years, the applications of the computer vision concepts and information communication technology has been observed in number of applications including home automation, healthcare, smart cities, precision agriculture etc. Internet of Things (IoT) is the underlying technology that indulges in almost all part of world infrastructure with the indispensable concept of connecting every device for collecting, contributing, experiencing, and analyzing the information. Smart or precision farming is known for achieving intelligence in agriculture. Therefore, in this article, an effort has been made towards automated disease detection from the plant leaves. For this a novel framework, a method named as IoT_FBFN using Fuzzy Based Function Network (FBFN) enabled with IoT has been proposed. At first, the images of leaf are acquired. Then these images are preprocessed and features are extracted using the Scale-invariant feature transform method. Finally, FBFN is used for the detection of the galls caused by the insect named as Pauropsyllatuberculate. The training process of the network is by optimizing with the help of Firefly algorithm, this increases the efficiency of the network. The proposed IoT_FBFN network having the computational power of fuzzy logic and learning adaptability of neural network achieves higher accuracy for identification and classification of galls when compared with the other approaches. The article concludes with the challenges encountered and future works. [ABSTRACT FROM AUTHOR]

Automated pavement distress detection benefits road maintenance and operation by providing the condition and location of various distress rapidly. Existing work generally relies on manual labor or specific algorithms trained by dedicated datasets, which hinders the efficiency and applicable scenarios of methods. Street view map provides interactive panoramas of a large scale of urban roadway network, and is updated in a recurrent manner by the provider. This paper proposed a deep learning method based on a pre-trained neural network architecture to identify and locate different distress in real-time. About 20,000 street view images were collected and labeled as the training dataset using the Baidu e-map. Eight types of distress are notated using Yolov3 deep learning architecture. The scale-invariant feature transform (SIFT) descriptors combined with GPS and bounding boxes were applied to judge the deterioration of the distress. A decision tree was designed to evaluate the change of the distress over some time. A typical road in Shanghai was selected to verify the effectiveness of the proposed model. The images of the road from 2015 to 2017 were collected from the street view map. The results showed that the mean average precision of the proposed algorithm is 88.37%, demonstrating the vast potential of applying this method to detect pavement distress. 43 distress were newly generated, and 49 previous distress were patched in the two years. The proposed method can assist the authorities to schedule the maintenance activities more effectively.

The advertising market's use of smartphones and kiosks for non-face-to-face ordering is growing. An advertising video recommender system is needed that continuously shows advertising videos that match a user's taste and displays other advertising videos quickly for unwanted advertisements. However, it is difficult to make a recommender system to identify users' dynamic preferences in real time. In this study, we propose an advertising video recommendation procedure based on computer vision and deep learning, which uses changes in users' facial expressions captured at every moment. Facial expressions represent a user's emotions toward advertisements. We can utilize facial expressions to find a user's dynamic preferences. For such a purpose, a CNN-based prediction model was developed to predict ratings, and a SIFT algorithm-based similarity model was developed to search for users with similar preferences in real time. To evaluate the proposed recommendation procedure, we experimented with food advertising videos. The experimental results show that the proposed procedure is superior to benchmark systems such as a random recommendation, an average rating approach, and a typical collaborative filtering approach in recommending advertising videos to both existing users and new users. From these results, we conclude that facial expressions are a critical factor for advertising video recommendations and are helpful in properly addressing the new user problem in existing recommender systems. [ABSTRACT FROM AUTHOR]

Breast cancer is considered the most serious cancer in women, among other prevalent cancer types. Chances of survival can be significantly increased through early detection, for which mammogram is considered to be the gold standard. This work addresses the development of a computer-aided diagnosis (CAD) system for analysis of mammographic masses. However, different mammographic tissue densities exhibit different characteristics and present abnormalities diversely. This renders a unified CAD framework for evaluation of masses, ineffective. To this end, we propose an ensemble framework for mass characterization, comprising different experts each specialized for a particular tissue category. Specifically, three segmentation-free feature descriptors including local binary pattern (LBP), scale-invariant feature transform (SIFT) and gray-level co-occurrence matrix (GLCM) are extracted from the regions of interest (ROIs), followed by individual classification with each feature descriptor using four different learning models, viz. support vector machine (SVM), artificial neural network (ANN), random forest (RF) and extreme grading boosting machine (XG Boost). All these 12 combinations are explored for each of the four breast density categories separately, to determine the best feature–classifier combination for a given category. The proposed ensemble scheme was validated on 1057 suspicious regions from digital database for screening mammograms (DDSM), demonstrating an improved performance when compared to state-of-the-art single learning framework modeled on all density categories collectively. [ABSTRACT FROM AUTHOR]

The pyramid Lucas-Kanade (LK) optical flow algorithm has been widely used in velocity measurement applications. However, these applications are limited by some shortcomings of the algorithm, such as its slow calculation speed and susceptibility to illumination changes. To solve these problems, a data fusion scheme based on the scale-invariant feature transform (SIFT) and optical flow is proposed to alleviate the dependence of the optical flow on the illumination conditions. In addition, an improved cubature Kalman filter (CKF) based on multi-rate residual correction (CKF-MRC) is proposed to solve the problem of inconsistency between the sampling frequencies of the SIFT and the optical flow, and takes full advantage of the high sampling frequency of SIFT. The experimental results demonstrate that the proposed CKF-MRC method can effectively improve the accuracy of velocity measurement under variable illumination conditions with a high sampling frequency.

Nowadays, the algorithms most used for object recognition are based on a constructed database or on training and learning processes with many samples, allowing robots to effectively perform object recognition. If objects in a home environment do not appear in a database, a system cannot recognize and segment items from household goods. In this study, we proposed an algorithm to reduce the processing complexity of object recognition that combines a depth image, object segmentation, and model construction with the GrabCut algorithm, and uses a hierarchical design for the segmentation of items. This algorithm uses the depth image to find the approximate locations and sizes of multiple objects in a coarse layer, then it uses GrabCut as a fine segmentation technology to segment the edges of objects and construct the models. First, we use the inputs of binocular vision to generate an anaglyph image, which is used as the base to perceive the environment's 3D information. At the same time, the too distant background is filtered, then histogram segmentation of the analysis image is used to partition each object. Next, GrabCut is used to find a convergent partition on the masking image to generate complete object edges. Finally, the scale-invariant feature transform (SIFT) is used for the extraction and recognition of feature points, and the database is updated. [ABSTRACT FROM AUTHOR]

When using drone-based aerial images for panoramic image generation, the unstableness of the shooting angle often deteriorates the quality of the resulting image. To prevent these polluting effects from affecting the stitching process, this study proposes deep learning-based outlier rejection schemes that apply the architecture of the generative adversarial network (GAN) to reduce the falsely estimated hypothesis relating to a transform produced by a given baseline method, such as the random sample consensus method (RANSAC). To organize the training dataset, we obtain rigid transforms to resample the images via the operation of RANSAC for the correspondences produced by the scale-invariant feature transform descriptors. In the proposed method, the discriminator of GAN makes a pre-judgment of whether the estimated target hypothesis sample produced by RANSAC is true or false, and it recalls the generator to confirm the authenticity of the discriminator’s inference by comparing the differences between the generated samples and the target sample. We have tested the proposed method for drone-based aerial images and some miscellaneous images. The proposed method has been shown to have relatively stable and good performances even in receiver-operated tough conditions.