A5108055798- Emotion and Mood Recognition
- Face and Expression Recognition
- Human Pose and Action Recognition
- Video Surveillance and Tracking Methods
- Image and Signal Denoising Methods
- Advanced Image and Video Retrieval Techniques
- Face recognition and analysis
- EEG and Brain-Computer Interfaces
- Blind Source Separation Techniques
- Advanced Vision and Imaging
- Medical Image Segmentation Techniques
- Advanced Data Compression Techniques
- Image Retrieval and Classification Techniques
- Anomaly Detection Techniques and Applications
- Speech and Audio Processing
- Advanced Neural Network Applications
- Gait Recognition and Analysis
- Hand Gesture Recognition Systems
- Image Processing Techniques and Applications
- Neural Networks and Applications
- Gaze Tracking and Assistive Technology
- Advanced Image Processing Techniques
- Remote-Sensing Image Classification
- Advanced Optical Imaging Technologies
- Gastrointestinal Bleeding Diagnosis and Treatment
Brunel University of London
2016-2025
University of London
2024-2025
Tianjin Chengjian University
2023-2024
Physical Sciences (United States)
2024
Chongqing University of Posts and Telecommunications
2022
Southwest University
2018-2019
Florida Institute of Technology
2017
London Centre for Nanotechnology
2011-2016
University College London
2011-2016
Apple (United States)
2014-2015
Recently, Multi-Object Tracking (MOT) has attracted rising attention, and accordingly, remarkable progresses have been achieved. However, the existing methods tend to use various basic models (e.g, detector embedding model), different training or inference tricks, etc. As a result, construction of good baseline for fair comparison is essential. In this paper, classic tracker, i.e., DeepSORT, first revisited, then significantly improved from multiple perspectives such as object detection,...
As fuzzy c-means clustering (FCM) algorithm is sensitive to noise, local spatial information often introduced an objective function improve the robustness of FCM for image segmentation. However, introduction leads a high computational complexity, arising out iterative calculation distance between pixels within neighbors and centers. To address this issue, improved based on morphological reconstruction membership filtering (FRFCM) that significantly faster more robust than proposed in paper....
Abstract Deep learning has been widely used for medical image segmentation and a large number of papers presented recording the success deep in field. A comprehensive thematic survey on using techniques is presented. This paper makes two original contributions. Firstly, compared to traditional surveys that directly divide literatures into many groups introduce detail each group, we classify currently popular according multi‐level structure from coarse fine. Secondly, this focuses supervised...
A great number of improved fuzzy c-means (FCM) clustering algorithms have been widely used for grayscale and color image segmentation. However, most them are time-consuming unable to provide desired segmentation results images due two reasons. The first one is that the incorporation local spatial information often causes a high computational complexity repeated distance computation between centers pixels within neighboring window. other regular window usually breaks up real structure thus...
Pain-related emotions are a major barrier to effective self rehabilitation in chronic pain. Automated coaching systems capable of detecting these potential solution. This paper lays the foundation for development such by making three contributions. First, through literature reviews, an overview how pain is expressed and motivation it physical provided. Second, fully labelled multimodal dataset (named `EmoPain') containing high resolution multiple-view face videos, head mounted room audio...
Clustering algorithms by minimizing an objective function share a clear drawback of having to set the number clusters manually. Although density peak clustering is able find clusters, it suffers from memory overflow when used for image segmentation because moderate-size usually includes large pixels leading huge similarity matrix. To address this issue, here we proposed automatic fuzzy framework (AFCF) segmentation. The has threefold contributions. First, idea superpixel (DP) algorithm,...
Depression is a typical mood disorder, and the persons who are often in this state face risk mental even physical problems. In recent years, there has therefore been increasing attention machine based depression analysis. such low mood, both facial expression voice of human beings appear different from ones normal states. This paper presents novel method, which comprehensively models visual vocal modalities, automatically predicts scale depression. On one hand, Motion History Histogram (MHH)...
A human being's cognitive system can be simulated by artificial intelligent systems. Machines and robots equipped with capability automatically recognize a humans mental state through their gestures facial expressions. In this paper, an is proposed to monitor depression. It predict the scales of Beck depression inventory II (BDI-II) from vocal visual First, different features are extracted expression images. Deep learning method utilized extract key frames. Second, spectral low-level...
The increasing number of people playing games on touch-screen mobile phones raises the question whether touch behaviors reflect players’ emotional states. This prospect would not only be a valuable evaluation indicator for game designers, but also real-time personalization experience. Psychology studies acted behavior show existence discriminative affective profiles. In this article, finger-stroke features during gameplay an iPod were extracted and their power analyzed. Machine learning...
Morphological reconstruction (MR) is often employed by seeded image segmentation algorithms such as watershed transform and power it able to filter seeds (regional minima) reduce over-segmentation. However, MR might mistakenly meaningful that are required for generating accurate also sensitive the scale because a single-scale structuring element employed. In this paper, novel adaptive morphological (AMR) operation proposed has three advantages. Firstly, AMR can adaptively useless while...
Nonuniform brightness distribution of collimator reticle images leads to the recognition difficulty grids and coordinate scales on affects digitalization system. The article proposes an image quality improvement algorithm that operates L channel LAB color space perform correction. It designs four typical modules depending representative features. Corresponding correction models for all are constructed by designing new functions based nonlinear gamma transform. A concept similarity is defined...
In this paper, we propose a human action recognition system suitable for embedded computer vision applications in security systems, human-computer interaction and intelligent environments. Our is application based on three reasons. Firstly, the was linear support vector machine (SVM) classifier where classification progress can be implemented easily quickly hardware. Secondly, use compacted motion features obtained from videos. We address limitations of well known history image (MHI) new...
Depression is a state of low mood and aversion to activity that can affect person's thoughts, behavior, feelings sense well-being. In such mood, both the facial expression voice appear different from ones in normal states. this paper, an automatic system proposed predict scales Beck Inventory naturalistic patients with depression. Firstly, features are extracted corresponding video audio signals represent characteristics vocal under Secondly, dynamic generation method feature space based on...
Automatic continuous affective state prediction from naturalistic facial expression is a very challenging research topic but important in human-computer interaction. One of the main challenges modeling dynamics that characterize expressions. In this paper, novel two-stage automatic system proposed to continuously predict dimension values videos. first stage, traditional regression methods are used classify each individual video frame, while second time-delay neural network (TDNN) model...
Human brain behavior is very complex and it difficult to interpret. emotion might come from activities. However, the relationship between human activities far clear. In recent years, more researchers are trying discover this by recording signals such as electroencephalogram (EEG) with associated information extracted other modalities facial expression. paper, machine learning based methods used model in publicly available dataset DEAP (Database for Emotional Analysis using Physiological...
Micro-expression is a subtle and involuntary facial expression that may reveal the hidden emotion of human beings. Spotting micro-expression means to locate moment when happens, which primary step for recognition. Previous work in spotting focus on from short video, with hand-crafted features. In this paper, we present methodology long videos. Specifically, new convolutional neural network named was designed extracting features video clips, first time deep learning used spotting. Then,...
Recently, real-time facial expression recognition has attracted more and research. In this study, an automatic system was built tested. Firstly, the model were designed tested on a MATLAB environment followed by Simulink that is capable of recognizing continuous expressions in with rate 1 frame per second implemented desktop PC. They have been evaluated public dataset, experimental results promising. The dataset labels used study made from videos, which recorded twice five participants while...
Traditional fuzzy clustering algorithms suffer from two problems in image segmentations. One is that these are sensitive to outliers due the non-sparsity of memberships. The other often cause over-segmentation loss local spatial information. To address issues, we propose a robust self-sparse algorithm (RSSFCA) for segmentation. proposed RSSFCA makes contributions. first concerns regularization under Gaussian metric integrated into objective function obtain membership with sparsity, which...
With the rapid development of deep learning (DL) in recent years, automatic modulation recognition (AMR) with DL has achieved high accuracy. However, aiming to obtain higher classification accuracy, requires numerous training samples. In order solve this problem, it is a challenge study how efficiently use for AMR case few letter, inspired by capsule network (CapsNet), we propose new structure named AMR-CapsNet achieve accuracy signals fewer samples, and further analyze adaptability models...
Naturalistic affective expressions change at a rate much slower than the typical which video or audio is recorded. This increases probability that consecutive recorded instants of represent same content. In this paper, we exploit such relationship to improve recognition performance continuous naturalistic expressions. Using datasets (AVEC 2011 and dataset, PAINFUL dataset) continuously labeled over time different dimensions, analyze transitions between levels those dimensions (e.g., in pain...