Blurriness is one of the common distortions in Images. This distortion is caused by spilling the pixel information overthe adjacent pixels. Blurriness has different types. The knowledge about the type of Image blurriness is one of the important parameters which directly affects performance of de-blurring methods. In this paper, a method has been proposed to classify the fourtypes of global blurrinessin digital Imagesin the spatial domain. These blurriness include the Gaussian blur, rectangular blur, motion blur and defocus blur. In the proposed method, the correlation concept is used to classify the type of Image blurriness. The correlation concept depicts the relations between the Image pixels. Also, the model and correlation of adjacent pixels are proportional to the type of blurriness. Appropriate features are extracted to detect the type of blurriness. The accuracy of the proposed method for detecting the type of blurriness is 90. 4%. This method has a better performance compared to other existing methods in terms of accuracy and computational cost.

Nowadays, with the spread of communication networks, the demand to transmit multimedia data has significantly increased. So, the knowledge about data type which is transmitted through the network is an important issue for monitoring communications and preventing transmission of malicious data. A typical identification system attempts to identify the type of transmitted coded data through Classification within a predefined set. The Classification is usually based on some relevant features extracted from the received bit stream. Most of the researches in this field consider a few kinds of Image codec in their Classification problem. In this paper, an efficient identification system is proposed for Classification within ten different Images codecs. The proposed system is based on combination and extension of existing methods. According to simulation results, Image codecs are classified with average accuracy of 88. 90%. Among various codecs, GIF and BMP have the highest accuracy of 99. 3% and 92. 5%, respectively. On the other hand, FLIF and WEBP have the lowest accuracy 83. 3% and 83. 6%, respectively.

Image processing is a method to perform some operations on an Image, in order to get an enhanced Image or to extract some useful information from it. The conventional Image processing algorithms cannot perform well in scenarios where the training Images (source domain) that are used to learn the model have a different distribution with test Images (target domain). Also, many real world applications suffer from a limited number of training labeled data and therefore benefit from the related available labeled datasets to train the model. In this way, since there is the distribution difference across the source and target domains (domain shift problem), the learned classifier on the training set might perform poorly on the test set. Transfer learning and domain adaptation are two outstanding solutions to tackle this challenge by employing available datasets, even with significant difference in distribution and properties, to transfer the knowledge from a related domain to the target domain. The main assumption in domain shift problem is that the marginal or the conditional distribution of the source and the target data is different. Distribution adaptation explicitly minimizes predefined distance measures to reduce the difference in the marginal distribution, conditional distribution, or both. In this paper, we address a challenging scenario in which the source and target domains are different in marginal distributions, and the target Images have no labeled data. Most prior works have explored two following learning strategies independently for adapting domains: feature matching and instance reweighting. In the instance reweighting approach, samples in the source data are weighted individually so that the distribution of the weighted source data is aligned to that of the target data. Then, a classifier is trained on the weighted source data. This approach can effectively eliminate unrelated source samples to the target data, but it would reduce the number of samples in adapted source data, which results in an increase in generalization errors of the trained classifier. Conversely, the feature-transform approach creates a feature map such that distributions of both datasets are aligned while both datasets are well distributed in the transformed feature space. In this paper, we show that both strategies are important and inevitable when the domain difference is substantially large. Our proposed using sample-oriented Domain Adaptation for Image Classification (DAIC) aims to reduce the domain difference by jointly matching the features and reweighting the instances across Images in a principled dimensionality reduction procedure, and construct new feature representation that is invariant to both the distribution difference and the irrelevant instances. We extend the nonlinear Bregman divergence to measure the difference in marginal, and integrate it with Fisher’ s linear discriminant analysis (FLDA) to construct feature representation that is effective and robust for substantial distribution difference. DAIC benefits pseudo labels of target data in an iterative manner to converge the model. We consider three types of cross-domain Image Classification data, which are widely used to evaluate the visual domain adaptation algorithms: object (Office+Caltech-256), face (PIE) and digit (USPS, MNIST). We use all three datasets prepared by and construct 34 cross-domain problems. The Office-Caltech-256 dataset is a benchmark dataset for cross-domain object recognition tasks, which contains 10 overlapping categories from following four domains: Amazon (A), Webcam (W), DSLR (D) and Caltech256 (C). Therefore 4 × 3 = 12 cross domain adaptation tasks are constructed, namely A → W, . . ., C → D. USPS (U) and MNIST (M) datasets are widely used in computer vision and pattern recognition tasks. We conduct two handwriting recognition tasks, i. e., usps-mnist and mnist-usps. PIE is a benchmark dataset for face detection task and has 41, 368 face Images of size 3232 from 68 individuals. The Images were taken by 13 synchronized cameras and 21 flashes, under varying poses, illuminations, and expressions. PIE dataset consists five subsets depending on the different poses as follows: PIE1 (C05, left pose), PIE2 (C07, upward pose), PIE3 (C09, downward pose), PIE4 (C27, frontal pose), PIE5 (C29, right pose). Thus, we can construct 20 cross domain problems, i. e., P1 → P2, P1 → P3, . . ., P5 → P4. We compare our proposed DAIC with two baseline machine learning methods, i. e., NN, Fisher linear discriminant analysis (FLDA) and nine state-of-the-art domain adaptation methods for Image Classification problems (TSL, DAM, TJM, FIDOS and LRSR). Due to these methods are considered as dimensionality reduction approaches, we train a classifier on the labeled training data (e. g., NN classifier), and then apply it on test data to predict the labels of the unlabeled target data. DAIC efficiently preserves and utilizes the specific information among the samples from different domains. The obtained results indicate that DAIC outperforms several state of-the-art adaptation methods even if the distribution difference is substantially large.

Semantic Classification of Images based on their low-level visual features is a challenging task in the field of Image retrieval and Classification. In this paper, the effect of weighting color, shape and texture feature vectors and also their components on Image Classification is investigated. The way that the Classification rate is affected by the database size is also studied. A database of 1000 Images from 10 semantic groups, 100 Images in each group, is used. A k-nearest neighbor classifier is employed and the leave-one-out rule is used to evaluate the results. The optimum weights for each type of the feature vectors and also their components are found by a genetic algorithm.  

In the past three decades, the use of smart methods in medical diagnostic systems has attracted the attention of many researchers. However, no smart activity has been provided in the  eld of medical Image processing for diagnosis of bladder cancer through cystoscopy Images despite the high prevalence in the world. In this paper, two well-known convolutional neural networks (CNNs) and a multilayer neural network was applied to classify bladder cystoscopy Images. One of the most im-portant issues in training phase of neural networks is determining the learning rate because selecting too small or large learning rate leads to slow convergence, volatility and divergence, respectively. Therefore, an algorithm is required to dynamically change the convergence rate. In this respect, an adaptive method was presented for determining the learning rate so that the multilayer neural network could be improved. In this method, the learning rate is determined using a coe cient based on the di erence between the accuracy of training and validation according to the output error. In addition, the rate of changes is updated according to the level of weight changes and output error. The proposed method was evaluated on 720 bladder cystoscopy Images in four classes of blood in urine, benign and malignant masses. Based on the simulated results, the second proposed method (CNNs) achieved at least 17% decrease in error and increased the convergence speed of the proposed method in the classi cation of cystoscopy Images, compared to the other competing methods.