نشریه ماشین بینایی و پردازش تصویر
سال نهم شماره 4 (زمستان 1401)

In this paper, we use the feature of fingerprint biometric signals to generate random cryptographic keys. The most important feature of this method is the use of unique fingerprint features in the production of random keys. The privacy of the randomly generated biometric key and the complexity of the key generation algorithm are key aspects of key security. In the proposed method, first the unique fingerprint features, which include Minosia dots, are extracted from the fingerprint image. Then, to increase the statistical properties and complexity, the Euclidean distance of all the points of Minosia relative to each other is calculated and stored in a matrix. In the next step, the data of this matrix is ​​moved to 8-bit numbers by permutation operation after normalization. Then, to increase the level of security and the ability to be random, the non-linear operation of the 8-bit S-box used in the AES format password is used. 8-bit data is applied to the S-box separately and the result is stored. Finally, data from low-byte to high-value bytes in 128-bit or 192-bit or 256-bit categories can be used as keys in cryptographic systems. Statistical analyzes performed on the generated keys show the acceptable randomness of the keys. Therefore, the proposed structure for generating a random key can be used in encrypting digital signals with large volumes of data such as image and sound.

In recent decades, various techniques in the field of computer vision have been proposed to classify and identify scenes in different spaces on visible spectrum images. Recently, improvements in various areas of machine vision using complementary features in visible and infrared images have been reported by researchers. This approach can also work for the subject of scene recognition. In this paper, we first create a multispectral image database, including pairs of color-visible and infrared spectrum images. Then, the images are fused by wavelet transform and using a weighting method based on training the deep convolutional neural networks. This approach is also compared with several other fusion methods using quantitative evaluation criteria. Finally, images of different scenes are classified using architectures based on deeply trained convolutional neural networks. Transfer learning approach is also used to classify the scene with the least computational cost. Experimental results show that the proposed method has been effective in scene classification, which is done by fusion of four-channel images (RGB-IR) with a classification accuracy of 96.67%. This method also shows better quantitative fusion criteria than other multispectral image fusion approaches.

Coronavirus 2019, or Covid-19, is an acute respiratory disease with high virus transmission capacity that has led to high mortality rates worldwide. Although rapid diagnosis can play an essential role in the patient's recovery, radiography by the treatment staff is a time-consuming process.Therefore, the use of ultrasound data and deep learning techniques are recommended. The ultrasound technique is radiation-free and can be used in pediatric wards and intensive care units for specific patients. However, its data have noise that affects the performance of deep learning methods. To this end, in this paper, we combine the uniform local noise-resistant binary pattern method with the deep learning method. First, a uniform local binary pattern is calculated on two temporal planes to extract the features of the Covid-19 manifestations in consecutive ultrasound data, and then, the resulting matrix is given as the input of the convolutional network. According to the performed experiments, the proposed method has better performance compared to that of the other state-of-art methods. The results show that the diagnosis accuracy of the Covid-19 from the ultrasound data using the proposed method is 98.5%.

Same family object matching is widely used in many applications such as object detection and tracking. In this article, we present an efficient way to search for objects in the same family. The proposed method is in fact an improved version of the BBS method, which in addition to having the positive features of this method such as resistance to background crowds, obstruction, scaling, change of angle and deformation, due to use of the characteristics of Haar wavelet, it is resistant to changes in light intensity and gives more accuracy in matching operation. The nature of the proposed method is based on the number of pairs of points that have the best matching, and has unique features that, by separating the features extracted from the image texture, make the matching operation be more accurate and less sensitive to changes in light intensity. Given the many difficulties in finding correspondence, we have proposed a method that overcomes many challenges. In the evaluation phase, by challenging the act of matching in a set of standard images of household objects, including the BBS database and other sports images, animals, objects, film scenes, and MRI images, the test results of the proposed method outperform against other methods.

Lip-reading or the process of visual recognition of speech is a way to distinguish speech from the movements of the speaker's lips. In recent years, the interest in automatic lip-reading has been increasing and a lot of researches have been done in this field and it is still a leading research topic. This paper provides an overview of researches in automatic lip-reading. A list of used databases in the researches is provided with specifications such as number of speakers, speech content and image quality. In this article, various lip-reading researches are reviewed from recognizing numbers, letters, words and sentences, and traditional methods of feature extraction and classification to the new methods based on deep learning. In addition, Persian lip-reading activities include provided databases and related researches are introduced more comprehensively.

Breast cancer is the most common type of cancer in the female population of the world. Early diagnosis and effective treatment with the aim of reducing mortality from this disease is done through screening methods.Ultrasound imaging is one of the most important and effective methods for the detection and diagnosis of this disease due to its non-invasive nature and its advantages over other diagnostic methods.Computer aided diagnosis systems were introduced to improve diagnostic performance. In this study, a fully automated breast cancer detection and diagnosis system is presented, which consists of four main steps: Image preprocessing in two steps to highlight the low echo areas (Hypo echo) and with the aim of selecting the seed point and the region of interest. As well as noise removal using non-local means filter, segmentation using image spatial and frequency information and improving it using genetic optimization algorithm, extracting a set containing 21 features based on shape and boundary Finally, the classification using the support vector machine classifier in order to separate the masses into benign and malignant groups. The results of evaluations performed on images of different datasets showed an accuracy of more than 95.5%.