A5015985183- Generative Adversarial Networks and Image Synthesis
- Advanced Image Processing Techniques
- Advanced Image and Video Retrieval Techniques
- Face recognition and analysis
- Image and Signal Denoising Methods
- Face and Expression Recognition
- Image Enhancement Techniques
- Advanced Neural Network Applications
- Aesthetic Perception and Analysis
- Emotion and Mood Recognition
- Advanced Vision and Imaging
- Advanced Image Fusion Techniques
- Brain Tumor Detection and Classification
- Image Retrieval and Classification Techniques
- Human Pose and Action Recognition
- Robotics and Sensor-Based Localization
- Maritime Navigation and Safety
- Computer Graphics and Visualization Techniques
- Smart Agriculture and AI
- Advanced Chemical Sensor Technologies
- IoT and Edge/Fog Computing
- Sparse and Compressive Sensing Techniques
- Simulation and Modeling Applications
- Educational Technology and Pedagogy
- Visual Attention and Saliency Detection
Northwest University
2012-2024
Northwest University
2024
State Administration of Cultural Heritage
2022
Northwestern Polytechnical University
2015-2020
Many recent 6D pose estimation methods exploited object 3D models to generate synthetic images for training because labels come free. However, due the domain shift of data distributions between real and images, network trained only on fails capture robust features in estimation. We propose solve this problem by making insensitive different domains, rather than taking more difficult route forcing be similar images. Inspired adaption methods, a Domain Adaptive Keypoints Detection Network...
Abstract Image restoration is a prominent field of research in computer vision. Restoring broken paintings, especially ancient Chinese artworks, significant challenge for current models. The difficulty lies realistically reinstating the intricate and delicate textures inherent original pieces. This process requires preserving unique style artistic characteristics paintings. To enhance effectiveness restoring traditional this paper presents framework called Sketch-Guided Restoration...
Automatic facial expression recognition (FER) plays an important role in many fields. However, most existing FER techniques are devoted to the tasks constrained conditions, which different from actual emotions. To simulate spontaneous expression, number of samples acted databases is usually small, limits ability classification. In this paper, a novel database for natural constructed leveraging social images and then deep model trained based on naturalistic dataset. An amount labeled obtained...
Ancient murals embody profound historical, cultural, scientific, and artistic values, yet many are afflicted with challenges such as pigment shedding or missing parts. While deep learning-based completion techniques have yielded remarkable results in restoring natural images, their application to damaged has been unsatisfactory due data shifts limited modeling efficacy. This paper proposes a novel progressive reasoning network designed specifically for mural image completion, inspired by the...
The recent significant progress in face recognition is mainly achieved using learning-based (LE) techniques via an exhaustive training involving a huge number of samples. However, many applications, the images available for may be very limited. This makes LE impractical learning discriminative features and models. Thus, limited samples (i.e. scarce data) degrades performance most existing methods. To overcome this problem, authors propose novel approach based on two-layer collaborative...
Similar to the basic facial expression recognition, one challenge for pain intensity recognition is some individual characteristics, e.g. face shapes, may cause great diversities in same emotion. So it usually very difficult distinguish two adjacent levels of as each has a large variation. In this study, coarse‐to‐fine combination method proposed recognition. The results multi‐scale outputs from multiple base deep network are combined probabilistic way improving discrimination between...
<title>Abstract</title> Existing image super-resolution methods have made remarkable advancements in enhancing the visual quality of real-world images. However, when it comes to restoring Chinese paintings, these encounter unique challenges. This is primarily due difficulty preserving intricate non-realistic details and capturing complex semantic information with high dimensionality. Moreover, preservation original artwork’s distinct style subtle artistic nuances fur- ther amplifies this...
Abstract How to migrate text-to-image models based on pre-trained diffusion adapt them domain generation tasks is a common problem. In particular, the task for Chinese landscape paintings with unique characteristics suffers from scarcity of fine-grained contextual details specific such artwork. Moreover, use substantial amounts non-landscape painting data during pre-training predisposes model be swayed by alternative visual styles, thereby leading generated images that inadvertently lack...
<title>Abstract</title> This paper addresses the challenging task of novel view synthesis for traditional Chinese landscape paintings, which typically offer only a single perspective and lack clear depth information. To overcome limitations existing methods that rely on multi-view input estimation, we propose method termed MVSM-CLP. The proposed CLPDepth Module employs high-low resolution fusion mechanism to enhance detail expression while preserving original scene structure. We introduce an...
Abstract Existing image super-resolution methods have made remarkable advancements in enhancing the visual quality of real-world images. However, when it comes to restoring Chinese paintings, these encounter unique challenges. This is primarily due difficulty preserving intricate non-realistic details and capturing comple semantic information with high dimensionality. Moreover, preservation original artwork’s distinct style subtle artistic nuances further amplifies this complexity. To...
Abstract Facial expressions in nonhuman primates are complex processes involving psychological, emotional, and physiological factors, may use subtle signals to communicate significant information. However, uncertainty surrounds the functional significance of facial animals. Using artificial intelligence (AI), this study found that exhibit undetectable by human observers. We focused on golden snub‐nosed monkeys ( Rhinopithecus roxellana ), a primate species with multilevel society. collected...
Han Chang'an City is the first cosmopolitan city and was largest metropolis at that time in world. As well known for center of nationality culture, has been considered as famous culture heritage plays an important role exploring 5000 years history China. With purpose reviving glories City, model reconstruction, navigation system user interaction with technique virtual scene were performed this paper. In system, Unity3D chosen 3D platform, Max used to scene, JavaScript C# are programming...
Spontaneous facial expression achieves much attention recently as it has potential applications in the field of computer vision and pattern recognition. Although convolutional networks have been applied for recognizing acted expressions obtained state-of-the-art performance, performance spontaneous still needs to be improved. In this paper, a heterogeneous deep model is presented recognize expressions. The consists two types with different architectures. To leverage data, these sub-networks...
Augmented reality (AR) aims to combine real and virtual worlds, offering real-time 3D interactions between users information. Existed AR systems detect specific markers in environments augmented them with contents such as objects videos. With the explosive growth of powerful, less expensive mobile devices, are extended devices. However, they fail achieve accurate integration world because insufficient detection accuracy computational resource shortage. In this paper, we propose a hybrid...
Calligraphy (the special art of drawing characters with a brush specially made by the Chinese) is an integral part Chinese culture, and detecting calligraphy highly significant. At present, there are still some challenges in detection ancient calligraphy. In this paper, we interested character problem focusing on boundary. We chose High-Resolution Net (HRNet) as feature extraction backbone network to learn reliable high-resolution representations. Then, used scale prediction branch spatial...