Scale Invariant Feature Transform Method Research Articles

Design of automatic style classification system of visual art works based on image processing

Abstract The emergence of a vast array of visual artworks meets the growing public demand for cultural literacy while simultaneously presenting challenges in the understanding and selection of appropriate artistic styles. Leveraging advancements in image processing technology, this study successfully automates the classification of 158,652 visual art images by categorizing them into distinct artistic styles. A novel approach combining the Scale-Invariant Feature Transform (SIFT) feature extraction algorithm with the spatial pyramid matching technique is introduced. This paper details the development and evaluation of a classification system tailored for the automatic categorization of visual artworks. Comparative testing results demonstrate that this innovative combination surpasses the traditional SIFT algorithm in both execution speed and classification accuracy. Specifically, the Support Vector Machine (SVM) based classification system, utilizing the enhanced SIFT method, achieves an accuracy rate exceeding 90% in identifying five distinct styles, including PHPS, NOS, IS, TF, and AAS. The implementation of this automatic classification system proves to be a practical tool in aiding individuals to select the visual artwork style that best suits their preferences.

An Improved SIFT Underwater Image Stitching Method

Underwater image stitching is a technique employed to seamlessly merge images with overlapping regions, creating a coherent underwater panorama. In recent years, extensive research efforts have been devoted to advancing image stitching methodologies for both terrestrial and underwater applications. However, existing image stitching methods, which do not utilize detector information, heavily rely on matching feature pairs and tend to underperform in situations where underwater images contain regions with blurred feature textures. To address this challenge, we present an improved scale-invariant feature transform (SIFT) underwater image stitching method. This method enables the stitching of underwater images with arbitrarily acquired images featuring blurred feature contours and that do not require any detector information. Specifically, we perform a coarse feature extraction between the reference and training images, and then we acquire the target image and perform an accurate feature extraction between the reference and target images. In the final stage, we propose an improved fade-in and fade-out fusion method to obtain a panoramic underwater image. The experimental results show that our proposed method demonstrates enhanced robustness, particularly in scenarios where detecting feature points is challenging, when compared to traditional SIFT methods. Additionally, our method achieves higher matching accuracy and produces higher-quality results in the stitching of underwater images.

Light weight convolutional neural network and low-dimensional images transformation approach for classification of thermal images

Thermal energy is emitted in the infrared range between X-ray and Gamma rays, which are invisible to the human eye. Thermal cameras can detect the temperature that arises due to the heat emitted by the objects in a non-contact way and transform it into an image. These images ensure to detection of objects regardless of ambient occlusion. Based on this problem, five different classification models were proposed within the scope of the study. New low-dimensional images were obtained by extracting the features of thermal images with HOG (Histogram Oriented of Gradients), LBP (Local Binary Pattern), SIFT (Scale Invariant Feature Transform), and GF (Gabor Filter) methods. These images are classified by a CNN (Convolutional Neural Network) model called LW-CNN (Light Weight CNN). Raw thermal images were classified with the LW-CNN model without pre-processing. In order to analyze the efficiency of the proposed models, the results were compared via the pre-trained VGG16 model. Three different datasets containing thermal images were used in classification processes. The highest classification accuracy was obtained from the LW-CNN model in the performance evaluations carried out on the three datasets. With this model, the classification accuracies obtained from the datasets are 98.58%, 95.56%, and 100%, respectively.

Fast and Robust Exudate Detection in Retinal Fundus Images Using Extreme Learning Machine Autoencoders and Modified KAZE Features.

Diabetic retinopathy(DR) is a health condition that affects the retinal blood vessels(BV) and arises in over half of people living with diabetes. Exudates(EX) are significant indications of DR. Early detection and treatment can prevent vision loss in many cases. EX detection is a challenging problem for ophthalmologists due to its different sizes and elevations as retinal fundus images frequently have irregular illumination and are poorly contrasting. Manual detection of EX is a time-consuming process to diagnose a mass number of diabetic patients. In the domain of signal processing, both SIFT (scale-invariant feature transform) and SURF (speed-up robust feature) methods are predominant in scale-invariant location retrieval and have shown a range of advantages. But, when extended to medical images with corresponding weak contrast between reference features and neighboring areas, these methods cannot differentiate significant features. Considering these, in this paper, a novel method is proposed based on modified KAZE features, which is an emerging technique to extract feature points and extreme learning machine autoencoders(ELMAE) for robust and fast localization of the EX in fundus images. The main stages of the proposed method are pre-processing, OD localization, dimensionality reduction using ELMAE, and EX localization. The proposed method is evaluated based on the freely accessible retinal database DIARETDB0, DIARETDB1, e-Ophtha, MESSIDOR, and local retinal database collected from Silchar Medical College and Hospital(SMCH). The sensitivity, specificity, and accuracy obtained by the proposed method are 96.5%, 96.4%, and 97%, respectively, with the processing time of 3.19 seconds per image. The results of this study are satisfactory with state-of-the-art methods. The results indicate that the approach taken can detect EX with less processing time and accurately from the fundus images.

Contrast threshold adaptive adjustment algorithm for remote sensing image matching

Remote sensing image processing has been widely used in environmental monitoring, terrain survey, military investigation, disaster early warning, and other fields. The remote sensing images matching is a key step, and the accuracy and real-time of the alignment have a great impact on these applications. Due to the high resolution and complexity of remote sensing images, scale-invariant feature transformation (SIFT) algorithm has the problems of high computational complexity and poor matching effect. Adaptive threshold adjustment algorithm has proposed in the remote sensing image matching, but minor changes in the contrast threshold can bring about drastic changes in the image matching quality, which will affect applications such as monitoring, measurement, and early warning. To improve the matching quality of a SIFT detector, an adaptive contrast threshold SIFT method based on image complexity calculation (CACT-SIFT) is proposed. The CACT-SIFT finds two contrast thresholds for the target image and the reference image, respectively, and achieves the target by minimizing the proposed criteria Cik. Experiments show that the method can be applied to the matching of remote sensing images. The information of reference images and target images can be detected at the same time, and the key points can be extracted in a robust way, with better accuracy and real-time accuracy of the alignment.

Performance of SURF and SIFT Keypoints for the Automated Differentiation of Abnormality in Chest Radiographs.

In this work, automated abnormality detection using keypoint information from Speeded-Up Robust feature (SURF) and Scale Invariant Feature Transform (SIFT) descriptors in chest Radiographic (CR) images is investigated and compared. Computerized image analysis using artificial intelligence is crucial to detect subtle and non-specific alterations of Tuberculosis (TB). For this, the healthy and TB CRs are subjected to lung field segmentation. SURF and SIFT keypoints are extracted from the segmented lung images. Statistical features from keypoints, its scale and orientation are computed. Discrimination of TB from healthy is performed using SVM. Results show that the SURF and SIFT methods are able to extract local keypoint information in CRs. Linear SVM is found to perform better with precision of 88.9% and AUC of 91% in TB detection for combined features. Hence, the application of keypoint techniques is found to have clinical relevance in the automated screening of non-specific TB abnormalities using CRs.

Exploring The Growth Level Of Pancreatic Tumor Using Highly Efficient Cnn Classifier

The pancreatic cancer (PC) has the lowermost survivability and ranks as the fourth prominent reason of cancer death and the rate of death is increasedby ever year. The major risk factors for the invasion of pancreatic cancer are smoking, intake of alcohol, diabetes mellitus and earlier pancreatitis also leads to the development of pancreatic cancer. The aim of this suggested method is to detect the PC, which is accomplished by using the image processing technique. In this paper, the CT image is considered as the input image which experiences the process of preprocessing to eliminate the presence of noise in the image and that has been accomplished by adaptive Weiner filter. After the preprocessing, the noise free image is then segmented by using the modified region grow model. SIFT (Scale invariant feature transform) method is used to extract the parameters of the pancreatic cancer and this extracted features are enhanced by PCA (principal component analysis) to optimize the features of the pancreatic CT image. The parameters of the image have been stimulated by relating the CNN (convolutional neural network) classifier. After classification, the classified image has been compared, thus the comparison is done only between the test data and the trained data to categorize the image as PC or non-PC. The entire process is then stimulated in the MATLAB and most recent technique is used for the performance estimation and so high accuracy is achieved.

2D object recognition: a comparative analysis of SIFT, SURF and ORB feature descriptors

Object recognition is a key research area in the field of image processing and computer vision, which recognizes the object in an image and provides a proper label. In the paper, three popular feature descriptor algorithms that are Scale Invariant Feature Transform (SIFT), Speeded Up Robust Feature (SURF) and Oriented Fast and Rotated BRIEF (ORB) are used for experimental work of an object recognition system. A comparison among these three descriptors is exhibited in the paper by determining them individually and with different combinations of these three methodologies. The amount of the features extracted using these feature extraction methods are further reduced using a feature selection (k-means clustering) and a dimensionality reduction method (Locality Preserving Projection). Various classifiers i.e. K-Nearest Neighbor, Naive Bayes, Decision Tree, and Random Forest are used to classify objects based on their similarity. The focus of this article is to present a study of the performance comparison among these three feature extraction methods, particularly when their combination derives in recognizing the object more efficiently. In this paper, the authors have presented a comparative analysis view among various feature descriptors algorithms and classification models for 2D object recognition. The Caltech-101 public dataset is considered in this article for experimental work. The experiment reveals that a hybridization of SIFT, SURF and ORB method with Random Forest classification model accomplishes the best results as compared to other state-of-the-art work. The comparative analysis has been presented in terms of recognition accuracy, True Positive Rate (TPR), False Positive Rate (FPR), and Area Under Curve (AUC) parameters.

Face Recognition based on Convolution Neural Network and Scale Invariant Feature Transform

Recently, Face Recognition (FR) has been received wide attention from both the research community and the cyber security industrial companies. Low accuracy of recognition is considered a main challenge when it comes to talking about employing the Artificial Intelligence (AI) for FR. In this work, the Scale Invariant Feature Transform (SIFT) and the Convolutional Neural Networks (CNN) feature extraction methods are utilized to build an AI based classifier. The CNN extracts features through both the convolutional and polling layers, while the SIFT extracts features depending on the scale space, directions, and histograms of points of interest. The features that are extracted by the CNN and the SIFT methods are used as an inputs for the KNN classifier. The experimental results with 400 test images of 40 persons, with 240 images are randomly chosen as training sets and 160 images from test sets, demonstrate in terms of accuracy, sensitivity, and error rate, that the CNN-based KNN classifier achieved better results when compared to the SIFT-based KNN classifier (accuracy = 97%, sensitivity = 93%, error rate = 3%).

Implementasi Random Forest Untuk Klasifikasi Motif Songket Palembang Berdasarkan SIFT

Indonesia memiliki berbagai warisan budaya tak benda salah satunya adalah kain songket. Kain songket memiliki banyak ragam sesuai ciri khas dari setiap daerah, khususnya songket Palembang. Kain songket Palembang memiliki keistimewaan dibandingkan songket dari daerah lain. Selain memiliki nilai sejarah, kain songket Palembang memiliki motif, mutu dan tingkat kerumitan yang tinggi dalam proses pembuatannya. Pada penelitian ini digunakan metode Random Forest untuk klasifikasi citra motif kain songket Palembang dengan mengunakan ekstraksi fitur Scale-Invariant Feature Transform (SIFT). Proses pembentukan fitur dengan metode SIFT melalui tahap scale space extrema detection, keypoint localization, orientation assignment, dan keypoint descriptor. Fitur yang dihasilkan digunakan untuk klasifikasi Random Forest. Citra motif songket yang digunakan pada penelitian ini sebanyak 115 citra dari setiap jenis motif, yaitu Bunga cina, Cantik Manis, dan Pulir. Pemilihan citra diambil dari 5 warna setiap motif songket Palembang. Data latih dan data uji yang digunakan masing-masing sebanyak 100 dan 15 untuk setiap motif Songket Palembang. Hasil pengujian menunjukkan bahwa metode SIFT dan Random Forest untuk klasifikasi citra motif kain Songket Palembang dapat memberikan akurasi yang cukup baik, dimana metode SIFT dan Random Forest mampu menghasilkan rata-rata overall accuracy 92,98%, per class accuracy 94,07%, presision 92,98%, dan recall 89,74%.

Combined pattern matching and feature tracking for Bohai Sea ice drift detection using Gaofen-4 imagery

ABSTRACT Sea ice is the main cause of winter marine disasters in the Bohai Sea, which is an important sea area in China. The sea ice in the Bohai Sea is easily deformed by strong water currents and rapid temperature changes. Consequently, low spatial and temporal resolution data cannot meet the needs of high-precision monitoring of the sea ice in the Bohai Sea. The geostationary satellite Gaofen-4 (GF-4) has the unique advantages of a high temporal resolution (20 s) and a high spatial resolution (50 m) and can be used to monitor sea ice drift in the Bohai Sea. We evaluate a new algorithm that combines the scale-invariant feature transform (SIFT) technique and the normalized cross-correlation (NCC) method to obtain accurate sea ice drift information. The interclass separability method is used to determine the optimal band combination (OBC), and the results show that the OBC for distinguishing sea ice from sea water is B2×(B1× B5)−1. The proposed method improves noise immunity and calculation efficiency and considers the characteristics of ice blocks. The proposed algorithm can also obtain more correctly matched feature points than the NCC method and approximately 10 times more correct points than the SIFT technique. The mean absolute error (MAE) of the velocity obtained with the proposed algorithm is 0.1615 m s−1 lower than those obtained with the SIFT and NCC methods, and the root mean square error of the velocity determination of the three algorithms is very small. The MAE of the ice movement direction obtained with the proposed algorithm is 0.605° lower than those obtained with the SIFT and NCC methods, and the MAE of a second data set is 12.378° better than that of the SIFT method. The motion of sea ice obtained by the proposed algorithm is close to the actual movement of the sea ice in the Bohai Sea.

An Image Registration Method for Multisource High-Resolution Remote Sensing Images for Earthquake Disaster Assessment.

For earthquake disaster assessment using remote sensing (RS), multisource image registration is an important step. However, severe earthquakes will increase the deformation between the remote sensing images acquired before and after the earthquakes on different platforms. Traditional image registration methods can hardly meet the requirements of accuracy and efficiency of image registration of post-earthquake RS images used for disaster assessment. Therefore, an improved image registration method was proposed for the registration of multisource high-resolution remote sensing images. The proposed method used the combination of the Shi_Tomasi corner detection algorithm and scale-invariant feature transform (SIFT) to detect tie points from image patches obtained by an image partition strategy considering geographic information constraints. Then, the random sample consensus (RANSAC) and greedy algorithms were employed to remove outliers and redundant matched tie points. Additionally, a pre-earthquake RS image database was constructed using pre-earthquake high-resolution RS images and used as the references for image registration. The performance of the proposed method was evaluated using three image pairs covering regions affected by severe earthquakes. It was shown that the proposed method provided higher accuracy, less running time, and more tie points with a more even distribution than the classic SIFT method and the SIFT method using the same image partitioning strategy.

A New Sample Consensus Based on Sparse Coding for Improved Matching of SIFT Features on Remote Sensing Images

In this article, a new method is proposed for feature matching of remote sensing images using sample consensus based on sparse coding (SCSC) to improve the image registration technique. To this end, scale-invariant feature transform (SIFT) features are used to select interesting points for image matching. The extracted points contain some differences and similarities in two images captured from the same area (but different in sensor resolution, azimuth, elevation, contrast, illumination, etc.); in such a case, similar points should be extracted and other dissimilar should be eliminated. In this article, we greatly improve the matching between two images using the SCSC through checking points altogether. Moreover, the proposed method is shown to have better results than standard alternative methods such as random sample consensus (RANSAC) when the number of feature points is too much or have noise. However, it should be noted that for a low-noise and distortion rate, the proposed method and the RANSAC yield similar results. In general, the proposed method using sparse coding achieves a higher correct match rate than the SIFT algorithm. In order to illustrate this issue, the proposed method is compared to other updated matching and registration methods based on the SIFT algorithm. The obtained results confirm the accuracy of this claim and show that the proposed algorithm is accurate between 0.48% and 7.68% rather than SVD-RANSAC, Hoge, Stone, Foroosh, Leprince, Nagashima, Guizar, Youkyung, Lowe, Preregistration, IS-SIFT, SPSA, Gong, Standard SIFT, IS-SIFT, UR-SIFT, Sourabh, and Han methods.

A Contactless Zero-Value Insulators Detection Method Based on Infrared Images Matching

A contactless method based on infrared image matching is presented for zero-value insulator detection on outdoor porcelain insulator string. An improved SIFT (scale-invariant feature transformation) method is developed to extract features and accomplish the pre-matching between the insulator string to be detected and the standard string in the image library. An improved RANSAC (random sampling consistency) algorithm is developed to remove mismatching points and achieve more accurate and faster detection. Firstly, an adaptive circle window is designed to improve the sensitivity of the SIFT operator on arc features extraction from insulators. Then, the dimension of the feature is decreased from128 to 32 by dividing the circle into 4 fan-shaped regions and descript the feature on 8 directions to achieve the lower computation burden and high accuracy. For the SIFT method and RANSAC method uses Euclidean distance to measure the similarity between features extracted from the string to be detected and the string in the standard library, mismatching may be caused. Spatial geometric features of the neighborhood around the feature point are used to construct the error function to improve the RANSAC method and remove the mismatching points. Faster matching is obtained by a threshold control for decreasing the number of data checking for the consensus set data models. Testing results show that the presented method is effective in the detection of zero-value resistances for outdoor insulator string. Compare with the SIFT method and SIFT + RANSAC method, the presented method has higher accuracy and faster detection speed.

A Comparative Analysis of RootSIFT and SIFT Methods for Drowsy Features Extraction

Abstract Drowsiness during driving is a very serious concern in the society. Thousands of human lives are affected due to drowsy driving. So a system that can detect drowsy drivers is necessary in this situation. This paper does a comparative analysis of the approaches called Scale Invariant Feature Transform(SIFT) and RootSIFT for drowsy features extraction. RootSIFT is an enhanced SIFT descriptor. SIFT has been the widely used technique for feature extraction due to its invariance to scale, rotation, illumination, viewpoint and translations. So the enhancement to SIFT to detect drowsy features has made an outcome more likely. The enhanced SIFT called RootSIFT achieves 93.55% accuracy, which is better than normal SIFT in extracting the drowsy features.

GPU-based parallel optimization for real-time scale-invariant feature transform in binocular visual registration

Scale-invariant feature transform (SIFT) is one of the widely used interest point features. It has been successfully applied in various computer vision algorithms like object detection, object tracking, robotic mapping, and large-scale image retrieval. Although SIFT descriptor is highly robust towards scale and rotation variations, the high computational complexity of the SIFT algorithm inhibits its use in applications demanding real-time response and in algorithms dealing with very large-scale databases. In order to be effective for image matching process in near real-time, the Compute Unified Device Architecture (CUDA) application programming interface of a graphics processing unit (GPU) is incorporated to speed up or improve the SIFT method. Experimental results show that the proposed GPU-based SIFT framework is suitable for image application in real time. It can improve the image matching process both in time and accuracy compared with conventional SIFT method.

Copy move forgery detection using key point localized super pixel based on texture features

The most important barrier in the image forensic is to ensue a forgery detection method such can detect the copied region which sustains rotation, scaling reflection, compressing or all. Traditional SIFT method is not good enough to yield good result. Matching accuracy is not good. In order to improve the accuracy in copy move forgery detection, this paper suggests a forgery detection method especially for copy move attack using Key Point Localized Super Pixel (KLSP). The proposed approach harmonizes both Super Pixel Segmentation using Lazy Random Walk (LRW) and Scale Invariant Feature Transform (SIFT) based key point extraction. The experimental result indicates the proposed KLSP approach achieves better performance than the previous well known approaches.

SIFT Detector Boosted by Adaptive Contrast Threshold to Improve Matching Robustness of Remote Sensing Panchromatic Images

Although scale invariant feature transform (SIFT) based algorithms have a wide range of applications in remote sensing panchromatic image matching, but minor changes in the contrast threshold can bring about drastic changes in the image matching quality. In order to effectively improve the matching quality of the SIFT detector, this paper proposes an adaptive contrast threshold-SIFT (ACT-SIFT) procedure. The ACT-SIFT method set forth herein harmoniously calculates two sought after contrast thresholds for the target and the reference images separately an objective achieved through minimizing certain proposed criterion. As an introductory step, the entropy in the keypoints (EiK) is defined as a control or balancing parameter being estimated in terms of the entropy of the image and the number of the keypoints. To that end, the candidate keypoints are first extracted from the scale space of SIFT method by applying the initial contrast threshold for both images. Then, the contrast threshold values are modified iteratively for both images to reach the needed values by minimizing the relative difference between the EiK of the target and the reference images. Next, in agreement with the threshold values obtained, the keypoints are extracted with the help of the SIFT keypoint detector. The correct matching pairs are also created using the matches acquired through the SIFT descriptor. The results obtained by applying our proposed approach promise boosted matching pairs in remote sensing image correspondences, for it extracts the keypoints in a robust manner by simultaneously checking the information content of the reference and the target images.

Implementation of Parallel Processing on Multi-Object Recognition System Software

Multi-object recognition software on Remote Controlled Weapon Station (RCWS) had been implemented in previous paper using Scale Invariant Feature Transform (SIFT) and Speeded Up Robust Features (SURF) methods, but the processing time in one cycle is quite slow so it is need to be optimized using parallel processing. In this paper, implementation of parallel processing on multi-object recognition software has been done on a multicore processor. The Openmp Application Programming Interface (API), C programming language, and Visual studio Integrated Development Environment (IDE) is used to implement the parallel processing in this paper. The parallel processing was implemented in the for loop of the matching process between the capturing object from the camera and the database under two conditions, i.e., the original of the for loop syntax and after optimization of the for loop syntax. Experiments have been done on the core processor i7-4790 @ 3.60Ghz, 8 GB DDR3 of memory, windows 8.1 os using two, four, six, and eight cores to recognize one, two, three and four objects at once using SIFT and SURF methods. Based on the experiments, it was found that the processing time in parallel is faster than sequential process, where the fastest of the processing time is obtained after optimization in the loop syntax, with the processing time in recognizing one to four objects using SIFT method is 927.13 ms (8 core), 1019.31 ms (6 core), 1190.72 ms (8 core), and 1283.05 ms (4 core), where the sequential processing time in recognizing one to four objects is 1067.35 ms, 1164.78 ms, 1352.93 ms, and 1497.35 ms, while the processing time in recognizing one to four objects using SURF method is 1157.13 ms (8 core), 1517.83 ms (6 core), 1572.14 ms (4 core), dan 1472.64 ms (6 core), where the sequential processing time in recognizing one to four objects is 5635.99 ms, 6268.47 ms, 3256.63 ms, dan 3883.78 ms.

A-RANSAC: Adaptive random sample consensus method in multimodal retinal image registration

Abstract In this paper, an adaptive Random Sample Consensus (A-RANSAC) method is proposed for multimodal retinal image registration. In this method, the features of two various images from images taken with different modalities such as FA (Fluorescein angiography) and RF (Red free) are extracted using a modified version of Scale Invariant Feature Transform method (SIFT) called SAR-SIFT which is originally used for Synthetic Aperture Radar images. Then, the matching performance between these images is enhanced using the proposed A-RANSAC. In the A-RANSAC method, the threshold value is chosen so that the Root Mean Square Error (RMSE) and the number of removed matches are optimized simultaneously. The efficiency of the proposed method has been investigated in other modes such as high resolution and low-quality retinal image registration in addition to multimodal registration. The simulation results on several retinal image datasets show that the proposed method improves the precision matching by 9.89% and rate of success by 25% on the average compared to the SAR-SIFT method.