Qin Jingyan

University of Science and Technology Beijing | USTB · Department of Environmental Engineering

Recent methods of learning latent representations in Domain Adaptation (DA) often entangle the learning of features and exploration of latent space into a unified process. However, these methods can cause a false alignment problem and do not generalise well to the alignment of distributions with large discrepancy. In this study, the authors propose...

Text detection in complex scene images is a challenging task for intelligent transportation. Existing scene text detection methods often adopt multi-scale feature learning strategies to extract informative feature representations for covering objects of various sizes. However, the sampling operation inherent in multi-scale feature generation can ea...

Learning a common latent embedding by aligning the latent spaces of cross-modal autoencoders is an effective strategy for Generalized Zero-Shot Classification (GZSC). However, due to the lack of fine-grained instance-wise annotations, it still easily suffer from the domain shift problem for the discrepancy between the visual representation of diver...

Learning a common latent embedding by aligning the latent spaces of cross-modal autoencoders is an effective strategy for Generalized Zero-Shot Classification (GZSC). However, due to the lack of fine-grained instance-wise annotations, it still easily suffer from the domain shift problem for the discrepancy between the visual representation of diver...

Text-based visual question answering (VQA) requires to read and understand text in an image to correctly answer a given question. However, most current methods simply add optical character recognition (OCR) tokens extracted from the image into the VQA model without considering contextual information of OCR tokens and mining the relationships betwee...

Class imbalance, as a phenomenon of asymmetry, has an adverse effect on the performance of most machine learning and overlap is another important factor that affects the classification performance of machine learning algorithms. This paper deals with the two factors simultaneously, addressing the class overlap under imbalanced distribution. In this...

Text-based visual question answering (VQA) requires to read and understand text in an image to correctly answer a given question. However, most current methods simply add optical character recognition (OCR) tokens extracted from the image into the VQA model without considering contextual information of OCR tokens and mining the relationships betwee...

Teaching–learning-based optimization (TLBO) algorithm, which simulates the process of teaching–learning in the classroom, has been studied by many researchers, and a number of experiments have shown that it has great performance in solving optimization problems. However, it has an inherent origin bias in teacher phase and may fall into local optima...

Colormap is a visualization tool to realize the mapping from data to colors. The data patterns can be revealed by color distribution, and the data details can be explored by the mapping. Most colormaps use a linear mapping between data and colors. When the data are unevenly distributed, most data samples are encoded in very few colors, so that the...

Answer selection with deep neural networks has been studied extensively owing to its strong capability of encoding semantic features. Most previous work has treated the candidates as independent individuals for ranking but ignored the relations between these. In this study, we propose a ranking method via partial ordering for answer selection, whic...

Thyroid disease exists across the whole world and many people are suffering from this disease. The diagnosis of thyroid disease is of great importance to human life. Although there are already some researches that introduces various methods for thyroid diagnosis and achieves good results, the performance of diagnosis still needs to be improved. The...

The paper explores the function and contents design ontology for the interaction design of the family agent through design prototypes and models. Except for the FBS (Function + Behavior + Structure) Design Model for the agent product function design, the paper proposes the CMR (Contents + Mental/Interaction Model + Relationship/Requirements) Design...

With the smart speakers agent of Home Service Robot represented by voice interaction, tangible user interface interaction and somatosensory interaction are widely present in family environment and serve multiple family members, the trustworthy AI stimulates the transition of the interaction form from passive interaction to proactive interaction, fi...

With the increasing complexity and variety of robotic agents in VR/AR/MR IoT and artificial intelligence and other intelligent environments, traditional interaction grammar and interaction design principles are facing more and more challenges. Based on the theory of embodied interaction and disembodied interaction in cognitive science, this paper c...

Resumes are critical for individuals to find jobs and for HR to select staffs. Almost all key events about career and demographic information are recorded on the resumes, which are commercial secrets. Too many valuable patterns can be explored from resumes. In this paper, we collect 372,829 Chinese resumes with complete attributes, such as income a...

With the development of Information and Communications Technology (ICT), smart TV is gradually becoming universal and penetrating daily life. Smart TV has a wide range of user groups, and the elderly is an important group. The special physiological and psychological characteristics of Chinese elderly highlight the usability problems of smart TV int...

Intangible cultural heritage (ICH) is a precious historical and cultural resource of a country. Protection and inheritance of ICH is important to the sustainable development of national culture. There are many different intangible cultural heritage items in China. With the development of information technology, ICH database resources were built by...

Autonomous vehicle extends human mobility time and space dimensions under the background of intelligent transportation system big data and artificial intelligence, and drives the three flows which include logistic, financial flow and information flow to expand the new human mobility interaction paradigm. Mobility includes not only space and physics...

This paper analyzes the human-robot interaction (HRI) design based on artificial intelligence technology. Combined with the needs of artificial intelligence technology, this paper summed up the new characteristics of human-robot interaction which include high dimension, high tolerance, complex scenario merged with context awareness computing, consc...

Recalling the previous information revolution, big data, cloud computing and the Internet of things has become a phenomenon characterized by obvious features, the bearing of technology has become an important support for economic development in the era of intelligence, mobile products are still the core media for people to access information. This...

As the number of documents grows larger and larger, it becomes more and more difficult for people to make sense of it all. People’s Daily is the official newspaper of Chinese Communist Party Central Committee, which has a decisive guiding role for the Chinese mainland politics in different periods. In this paper, we develop an interactive visual an...

This paper presents a system for editing strip-like patches on the mesh using swept surface. The user first selects a rough cylinder-like region, the boundary of which is automatically refined. A swept surface approximation is automatically done, including extraction of the trajectory and the corresponding deforming 2D curves. The swept surface is...

The core of innovation design for product service system is to provide products and services meeting the needs of users. Traditional user research methods have many drawbacks, such as data interference, fuzzy feedback, single-dimensional indicators lack mutual verification, these issues have not been effectively resolved. If the designer can’t unde...

Face recognition is gaining great significance. The last step of face recognition can be divided into face verification and face classification. Face verification algorithms have been well-developed in recent years, but face classification researches are still facing problems. For example, when training face classification models like Support Vecto...

The hypothesis of this paper is the examination of dome constructions using the space technology nanomaterial silica aerogel as insulating material. Silica aerogels are nanomaterials with excellent thermal insulations properties because 99.98% of their volume can be pure air. Insulation's origin is the Latin insula 'island', an " isolated land " we...

The hypothesis of this paper is the examination of dome constructions using the space technology nanomaterial silica aerogel as insulating material. Silica aerogels are nanomaterials with excellent thermal insulations properties because 99.98% of their volume can be pure air. Insulation's origin is the Latin insula 'island', an " isolated land " we...

Green IT products collect personal small data by wearable computing and Quantified Self, Internet of Things (IoT) and Internet of Everything (IoE) link the small data with the big data, and transform the user experience from User-Centered Design (UCD) into User Participated (Involved or engagement) Design (UPD). To collect the seamless and sustaina...

Face recognition is a long-standing challenging topic in computer science, especially on insufficient datasets. The obstacle also lies in the balance of speed and accuracy. Recently, many algorithms claim that they have obtained great performance with high accuracy, but they are not enough for real-time application. In this work, a novel fast and a...

Color transfer is to alter an image’s color composition by reference to the color characteristics of another image. In this paper, we build a system called artistic coloring that realizes automatic color transfer from famous paintings. It properly extracts the wonderful color characteristics of famous paintings and applies them to color transfer. S...

This paper aims at responding to the questions in the traditional education and promoting a new natural interface and interaction form for new e-learning. This paper combined new e-learning with traditional oil painting to create a museum installation based on eye tracking and computer image captured technology. As the new e-learning installation w...

This paper aims at responding to the questions in the traditional cultural heritage protection and exploring new visual representation forms and human-computer natural interaction design methods to carry forward the digital cultural heritage(DCH) projects. This paper combined interaction design with Chinese traditional shadow puppet art to create a...

The paper analyzes ¿Affordance¿ in Human-Computer Interaction design and stresses the idea of interdependency between the environment (object) and its user. The paper introduces TUI (Tangible User Interface) into interactive product design methods attempting to break the barrier set up by the traditional GUI metaphor. Finally through an TUI interac...