A5015792118- Optical Coherence Tomography Applications
- Cutaneous Melanoma Detection and Management
- Advanced Steganography and Watermarking Techniques
- Anomaly Detection Techniques and Applications
- Acoustic Wave Resonator Technologies
- Ultrasonics and Acoustic Wave Propagation
- Context-Aware Activity Recognition Systems
- Advanced Neural Network Applications
- Industrial Vision Systems and Defect Detection
- Chaos-based Image/Signal Encryption
- Topic Modeling
- Image and Signal Denoising Methods
- Software System Performance and Reliability
- Structural Health Monitoring Techniques
- Vibration and Dynamic Analysis
- Human Pose and Action Recognition
- Digital Media Forensic Detection
- Vibration Control and Rheological Fluids
- Algorithms and Data Compression
- Computational and Text Analysis Methods
- Advanced Text Analysis Techniques
- Advanced Image and Video Retrieval Techniques
- Geophysical Methods and Applications
- Advanced Data Compression Techniques
- Network Security and Intrusion Detection
Shanghai Institute of Technology
2025
Multimedia University
2012-2021
University of Nottingham
2009-2011
Behavioral changes in daily home activities may be linked with health problems. Therefore, anomaly detection on sequence pattern of is important for healthcare monitoring. In this paper, an method based Long Short-Term Memory (LSTM) neural network proposed to detect anomalies activities. A comparison study LSTM and Hidden Markov Model (HMM) was conducted evaluate their performance under different training set size model's hyperparameters. The experimental results demonstrated that comparable...
This paper is concerned with modeling ultrasonic wave propagation in epoxy materials to better understand NDE procedures and provide reliable input more complex models of guided layered structures. Different physical are considered the context how well they simulate (known) linear relationship between bulk attenuation coefficients frequency. The identified then extended mechanical properties, which vary gradually spatial dimension. achieved using electric circuit transmission line analogs...
This article provides an improved automated skin lesion segmentation method for dermoscopic images.There are several stages this method.These include the pre-processing steps such as resizing images and eliminating noise.Hair was removed reflective light reduced using morphological operations a median filter.The single green channel rescaled into new intensities, it provided highest accuracy.The threshold value calculated to separate region from healthy skin.Morphological were implemented...
This paper proposes an efficient way to effectively segment malignant melanoma in color dermoscopy images. A combination of methods are used the proposed technique, including smoothing filters, PSNR, Spline, edge detection, morphological operations and segmentation. The pre-processing step eliminates noise, smoothes image employs spline function improve while lesion from image. Manual boundary selection is as benchmark test accuracy automatic by algorithm. evaluation results show that method...
<span>The population of elderly is growing and projected to outnumber the youth in future. Many researches on assisted living technology were carried out. One focus areas activity monitoring elderly. AReM dataset a time series recognition for seven different types activities, which are bending 1, 2, cycling, lying, sitting, standing walking. In original paper, author used many-to-many Recurrent Neural Network recognition. Here, we introduced classification method where Gated Units with...
Bulk wave attenuation in epoxies has been found experimentally to be a linear function of frequency. Anecdotal evidence suggests that the formulations used for bulk input computational propagation models, including models guided waves, can take many forms: Either simple relationship between and frequency directly, or any number analytic could, principle, – examples being Maxwell, Kelvin-Voigt Zener anelastic solid hysteretic damping model sometimes employed studies structural vibrations....
The rapid rise and widespread of ‘Fake News’ has severe implications in the society today. Much efforts have been directed towards development methods to verify news reliability on Internet recent years. In this paper, an automated evaluation system was proposed. utilizes term several Natural Language Processing (NLP) techniques such as Term Frequency-Inverse Document Frequency (TF-IDF), Phrase Detection Cosine Similarity tandem with Latent Semantic Analysis (LSA). A collection 9203 labelled...
This paper discusses a data hiding application using steganography. The purpose of this is to introduce two phase steganography that allows the user first compress information into ASCII form our novel compression technique (compression utilizing reference points coding) then embed carrier least significant bit (LSB) algorithms. An additional function method it will also allow choose whether in lossy or lossless format before embedding carrier. flexibility hide with customizable phase. By...
Network downtime is one of the most widely shared phenomenon within telecommunications infrastructure. In particular, faults from network equipment have received attention. Proactive monitoring system presented in this paper to address earliest symptoms malfunctioning equipment. Research focus has been placed on learning network's behavior as well detecting deviations MSAN (Multi-Service Access Node) norm at access layer. Additionally, aims provide an overview handling equipment, warnings,...
This paper proposed the reference points coding method for reversible data embedding. The provides user with flexibility to choose level of secret embedding from light embedded heavier depending on either size a certain image or that secrecy required during transfer. is achieved by first comparing each pixel value its neighboring pixels, when same values are found, own will be replaced two bit address pixel. remaining space then ready may embed up three bits per available while maintaining...
TensorFlow Object Detection API is an open-source object detection machine learning program that has gained recent popularity and being used in a variety of applications. Region-Based Fully Convolutional Network (R-FCN) Faster Neural (Faster R-CNN) are two models the very popular detection. This paper compares responses 2 when trained tested under same datasets for potholes. The compared their results evaluating superimposed with simple additive noises such as impulse noise, Gaussian noise...
This paper proposes a steganographic method that would improve transfer security and allow user customizable options, with no size restrictions for the secret image being transferred. It compares multiple directionally scanned Least Significant Bits of cover similarity in varied length segments. The result comparison will form basis generating key. A segment 2, 4, 6, 8 10 bits binary was used this to assess performance method. proposed led produce very good consistency results different...
This technical report summarizes submissions and compiles from Actor-Action video classification challenge held as a final project in CSC 249/449 Machine Vision course (Spring 2020) at University of Rochester