Search

Label distribution learning methods effectively address the label ambiguity problem and have achieved great success in image emotion analysis. However, these methods ignore structured and sparse information naturally contained in the annotations of emotions. For example, emotions can be grouped and ordered due to their polarities and degrees. Meanwhile, emotions have the character of intensity and are reflected in different levels of sparse annotations. Motivated by these observations, we present a convolutional neural network based framework called Structured and Sparse annotations for image emotion Distribution Learning (SSDL) to tackle two challenges. In order to utilize structured annotations, the Earth Mover’s Distance is employed to calculate the minimal cost required to transform one distribution to another for ordered emotions and emotion groups. Combined with Kullback-Leibler divergence, we design the loss to penalize the mispredictions according to the dissimilarities of same emotions and different emotions simultaneously. Moreover, in order to handle sparse annotations, sparse regularization based on emotional intensity is adopted. Through combined loss and sparse regularization, SSDL could effectively leverage structured and sparse annotations for predicting emotion distribution. Experiment results demonstrate that our proposed SSDL significantly outperforms the state-of-the-art methods.

In recent years, image style transfer has been greatly improved by using deep learning technology. However, when directly applied to clothing style transfer, the current methods cannot allow the users to self-control the local transfer position of an image, such as separating specific T-shirt or trousers from a figure, and cannot achieve the perfect preservation of clothing shape. Therefore, this paper proposes an interactive image localized style transfer method especially for clothes. We introduce additional image called outline image, which is extracted from content image by interactive algorithm. The interaction consists simply of dragging a rectangle around the desired clothing. Then, we introduce an outline loss function based on distance transform of the outline image, which can achieve the perfect preservation of clothing shape. In order to smooth and denoise the boundary region, total variation regularization is employed. The proposed method constrains that the new style is generated only in the desired clothing part rather than the whole image including background. Therefore, in our new generated images, the original clothing shape can be reserved perfectly. Experiment results show impressive generated clothing images and demonstrate that this is a good approach to design clothes.

As an information carrier with rich semantics, image plays an increasingly important role in real-time monitoring of logistics management. Abnormal objects are typically closely related to the specific region. Detecting abnormal objects in the specific region is conducive to improving the accuracy of detection and analysis, thereby improving the level of logistics management. Motivated by these observations, we design the method called abnormal object detection in a specific region based on Mask R-convolutional neural network: Abnormal Object Detection in Specific Region. In this method, the initial instance segmentation model is obtained by the traditional Mask R-convolutional neural network method, then the region overlap of the specific region is calculated and the overlapping ratio of each instance is determined, and these two parts of information are fused to predict the exceptional object. Finally, the abnormal object is restored and detected in the original image. Experimental results demonstrate that our proposed Abnormal Object Detection in Specific Region can effectively identify abnormal objects in a specific region and significantly outperforms the state-of-the-art methods.

Recently, automatic visual data understanding from drone platforms becomes highly demanding. To facilitate the study, the Vision Meets Drone Object Detection in Image Challenge is held the second time in conjunction with the 17-th International Conference on Computer Vision (ICCV 2019), focuses on image object detection on drones. Results of 33 object detection algorithms are presented. For each participating detector, a short description is provided in the appendix. Our goal is to advance the state-of-the-art detection algorithms and provide a comprehensive evaluation platform for them. The evaluation protocol of the VisDrone-DET2019 Challenge and the comparison results of all the submitted detectors on the released dataset are publicly available at the website: http: //www.aiskyeye.com/. The results demonstrate that there still remains a large room for improvement for object detection algorithms on drones.

As an information carrier with rich semantics, images contain more sentiment than texts and audios. So, images are increasingly used by people to express their opinions and sentiments in social network. The sentiments of the images are overall and should come from different regions. So, the recognition of the sentiment regions will help to concentrate on important factors the affect the sentiments. Meanwhile, deep learning method for image sentiment classification needs simple and efficient approach for simultaneously carrying out pruning and feature selection whilst optimizing the weights. Motivated by these observations, we design a region-based convolutional neural network using group sparse regularization for image sentiment classification: R-CNNGSR. The method obtains the initial sentiment prediction model through CNN using group sparse regularization to get compact neural network, and then automatically detect the sentiment regions by combining the underlying features and sentimental features. Finally, the whole image and the sentiment region are fused to predict the overall sentiment of the images. Experiment results demonstrate that our proposed R-CNNGSR significantly outperforms the state-of-the-art methods in image sentiment classification.

Video object detection has drawn great attention recently. The Vision Meets Drone Object Detection in Video Challenge 2019 (VisDrone-VID2019) is held to advance the state-of-the-art in video object detection for videos captured by drones. Specifically, there are 13 teams participating the challenge. We also report the results of 6 state-of-the-art detectors on the collected dataset. A short description is provided in the appendix for each participating detector. We present the analysis and discussion of the challenge results. Both the dataset and the challenge results are publicly available at the challenge website: http://www.aiskyeye.com/.

Current state-of-the-art image segmentation methods are mainly based on visible spectrum images. However, it remains challenging under adverse environmental conditions (e.g., nighttime, foggy, snowy). Thermal images can be captured in darkness and deployed in any weather and time. But the research focus on thermal images semantic segmentation remains underexploited since the lack of thermal images segmentation dataset of nighttime driving scene. On the other hand, existing methods do not adequately capture context between related pixels and edge details in thermal images. In this work, we focus on thermal image semantic segmentation of nighttime driving scene. First, we propose a multi-level correction network (MCNet) with a multi-level attention module (MAM) and a multi-level edge enhancement module (MEEM). Specifically, MAM selectively captures the inter-class and intra-class contextual dependencies by a multi-level correction process. In addition, MEEM continuously combines precise context information and edge prior knowledge in each level to correct the final feature representation. Second, we present a new dataset called SCUT-Seg which contains 2010 thermal images recorded from various road scenes with 10 manually annotated semantic region labels. Extensive experiments demonstrate that the proposed method performs favorably against the state-of-the-art methods on SCUT-Seg and the public MFNet dataset. The source code and dataset are available at https://github.com/haitaobiyao/MCNet .

Early nutrition is key to promoting gut growth and education of the immune system. Although iron deficiency anemia has long been recognized as a serious iron disorder, the effects of iron supplementation on gut development are less clear. Therefore, using suckling piglets as the model for iron deficiency, we assessed the impacts of iron supplementation on hematological status, gut development, and immunity improvement. Piglets were parenterally supplied with iron dextran (FeDex, 60 mg Fe/kg) by intramuscular administration on the third day after birth and slaughtered at the age of two days, five days, 10 days, and 20 days. It was expected that iron supplementation with FeDex improved the iron status with higher levels of serum iron, ferritin, transferrin, and iron loading in the liver by regulating the interaction of hepcidin and ferroportin (FPN). FeDex supplementation increased villus length and crypt depth, attenuated the pathological status of the duodenum, and was beneficial to intestinal mucosa. FeDex also influenced the intestinal immune development by stimulating the cytokines&rsquo, production of the intestine and enhancing the phagocytotic capacity of monocytes. Overall, the present study suggested that iron supplementation helped promote the development of the intestine by improving its morphology, which maintains its mucosal integrity and enhances the expression of immuno-associated factors.

Class overlapping and ensemble learning method have long been regarded as one of the toughest yet pervasive problems in data mining. The effect of ensemble learning method and class overlapping are with few researches. To meet this critical challenge, in this paper, we pay a systematic study on the different classifier performance in class overlapping problem and propose a new Ensemble Learning Method based on Naive Bayes for class overlapping problem (ELM-NB). ELM-NB uses NB to find class overlapping region and use this region and non-overlapping region in ensemble learning model. Experimental results on selected data sets prove that ELM-NB can have good classification performance for class overlapping problem.

Fraud in consumer-to-consumer (C2C) e-commerce is becoming more and more serious. The purpose of this study is to develop an effective fraud detection model to assist customers in identifying potential fraud transactions. We use Naive Bayes (NB), decision tree C4.5 and AdaBoost to construct the model for classifying imbalance transaction data, and majority voting is used to combine the model. Several experiments are conducted on Taobao data set to verify the classification performance of the proposed model using four popular performance metrics. The experimental results demonstrate that the model based on NB and AdaBoost&C4.5 can significantly increase the ability to locate potential fraud transactions in C2C e-commerce.

Using the internet for online information search has become the current main channel of getting information when tourists design a travel plan. However, with the rise of the internet and e-commerce websites, tourists are often submerged in a lot of information search and product selection. According to the customer service information of travel website, design a personalized travel package recommendation approach based on TD-LDA Model. The method uses topic analysis to extract the activities which tourists may interested in. And give full consideration to the duration of travel, the new package generated by travel company and other information to recommend personalized travel packages for tourists. Experimental results on real-world online customer service data show that this method can effectively extract the interests of tourists for travel package recommendation.

Iron deficiency is common throughout the world and has been linked to immunity impairments. Using piglets to model human infants, we assessed the impact of systemic iron homeostasis on proinflammatory status. Artificially reared piglets were parenterally supplied with iron dextran by intramuscular administration at the age of 3 days. Relative to no iron supplementation (control), iron dextran-treated (FeDex) piglets increased hematological parameters as well as iron levels in serum and tissues from days 21 to 49. High expression of hepcidin was observed in FeDex-treated piglets, which correlated with suppressed expression of ferroportin in duodenum. Lower levels of proinflammatory cytokine (IL-6, TNF-α, IFN-γ, and IL-1β) transcripts were detected in ileum of FeDex-treated piglets, which indicated that iron supplementation could attenuate the increase of inflammatory cytokines caused by iron deficiency. Histopathological analysis of liver and duodenum proved the less inflammatory responses after iron supplementation. Hepcidin was highly stimulated by FeDex supplementation and attenuated the inflammation of anemia, which implied that hepcidin might had antiinflammatory function and is a candidate regulator of the cross-talk between iron regulation and inflammation.

Successful segmentation of chromosomes is important for their analysis. Isolation of chromosomes is the major topic of the segmentation in previous studies, but it is at low resolution. While the model-based neural network (MBNN), in this paper, is adopted to segment the G bands of Triticum monococcum chromosomes, which is a high-resolution approach. Combined with the MBNN as a core technique, various auxiliary techniques have been used to find out the optimal method. A series of experimental results indicate that the optimal method for segmentation of G bands, here termed as MBNN-3P, is that using the MBNN as a core technique simultaneously aided by all of the three auxiliary techniques. They are presegmenting by the threshold T = 173, preassigning a class number and providing teacher's information. Therefore, it is feasible to apply the MBNN-3P to segment the G band images of T. monococcum chromosomes. This study might be of great significance to improve the accuracy and speed of automated analysis of plant chromosomes.

Development of an information system is a complex process, which expose to a great number of risks.Hence, the high failure rates long associated with information system projects, despite advances in techniques for information technology development, suggest that organizations need to improve their ability to identify and to manage associated risks. To improve the risk management in information system development projects, a pragmatic procedure is suggested to determine the size of a project's contingency plan budget at any specified level of certainty. Considering the interaction among risk factors, a method based on common risk factors and copula functions is used to model and quantify positive dependence between risks. DOI : http://dx.doi.org/10.11591/telkomnika.v12i6.5438 Full Text: PDF

Predicting products a customer would like on the basis of other customers ratings for these products has become a well-known approach adopted by many personalized recommendation systems on the Internet. With the development of electronic commerce, the number of customers and products grows rapidly, resulted in the sparsely of the rating dataset. Poor quality is one major challenge in collaborative filtering recommender systems. To solve this problem, slope one algorithm has a good performance. But there is also some places that not make sense. Slope one algorithm assumes that all users are at the same level, and it has neglected the individual differences. So, we propose an improved slope one algorithm, which will consider the weights of the user. Finally, we experimentally evaluate our approach and compare it to the original Slope One. The experiment shows that our method provides better recommendation results than it.

We propose a low-complexity single carrier frequency-domain equalizer and diversity combining for cooperative systems with demodulate-and-forward relaying over frequency-selective channels. At the relays, we apply the conventional frequency-domain linear equalizers and compute the output signal to noise ratios, which are used to calculate the normalized correlation coefficient for deriving an equivalent source-to-relay-destination (S-R-D) channel. At the destination, we propose the joint equalization and diversity combining receiver for symbol detection by utilizing the equivalent S-R-D channel. Finally, we demonstrate the superiority of the proposed scheme over the straightforward solution that ignores the decisions errors at the intermediate relays through simulations.

A method based on the model based neural network (MBNN) is proposed to automatically analyze high-resolution G bands of Triticum monococcum chromosomes. The MBNN-3P is employed to segmented the G-band images. Then five features of chromosomes are extracted from the segmented images. At last, the features are input into the MBNN for classification. The results indicate that the method provides a significant way for the automatic analysis of high-resolution G bands of plant chromosomes accurately.

The paper studied the small-scale multipath propagation characteristics of the wireless channel at 3.5 GHz, in a typical outdoor urban environment. A channel measurement platform was developed for the measurement. Both the fixed and mobile scenarios were measured; the power delay profile, the tapped delay line model, the multipath distribution, the mean excess delay and the Ricean K-factor were presented in the paper. The results verify that it is a rich scattered environment fit for multiple antenna wireless communications.

Contextual information and a priori knowledge play important roles in image segmentation based on neural networks. This paper proposed a method for including contextual information in a model-based neural network (MBNN) that has the advantage of combining a priori knowledge. This is achieved by including Markov random field (MRF) into the MBNN and this novel neural network is termed as MRF-MBNN. Then the proposed method is applied to segmenting the images. Experimental results indicate the MRF-MBNN is superior to the MBNN in image segmentation. This study is a successful attempt of incorporating contextual information and a prior knowledge into neural networks to segment images.