A5020849219- Image Enhancement Techniques
- Visual Attention and Saliency Detection
- Advanced Vision and Imaging
- Image and Video Quality Assessment
- Indoor and Outdoor Localization Technologies
- Speech and Audio Processing
- Ear and Head Tumors
- Business Process Modeling and Analysis
- Music and Audio Processing
- IoT-based Smart Home Systems
- Advanced Image and Video Retrieval Techniques
- Image Processing Techniques and Applications
- Flexible and Reconfigurable Manufacturing Systems
- Distributed and Parallel Computing Systems
- Energy Efficient Wireless Sensor Networks
- Head and Neck Cancer Studies
- Advanced Image Fusion Techniques
- Fire Detection and Safety Systems
- Underwater Vehicles and Communication Systems
- Mobile Ad Hoc Networks
- Biometric Identification and Security
- Distributed Sensor Networks and Detection Algorithms
- Service-Oriented Architecture and Web Services
- Neural Networks and Applications
- Industrial Vision Systems and Defect Detection
University of Warwick
2022-2024
University of Chicago
2023
Laboratoire des signaux et systèmes
2020-2021
Microsoft Research (United Kingdom)
2020
Microsoft (United States)
2019
Saarland University
2018
Stony Brook University
2011
Motorola (United States)
2007
We introduce phi-3-mini, a 3.8 billion parameter language model trained on 3.3 trillion tokens, whose overall performance, as measured by both academic benchmarks and internal testing, rivals that of models such Mixtral 8x7B GPT-3.5 (e.g., phi-3-mini achieves 69% MMLU 8.38 MT-bench), despite being small enough to be deployed phone. The innovation lies entirely in our dataset for training, scaled-up version the one used phi-2, composed heavily filtered web data synthetic data. is also further...
We consider the problem of localizing a wireless client in an indoor environment based on signal strength its transmitted packets as received stationary sniffers or access points. Several state-of-the-art localization techniques have drawback that they rely extensively labor-intensive 'training' phase does not scale well. Use unmodeled hardware with heterogeneous power levels further reduces accuracy these techniques.
We present a novel approach to named entity recognition (NER) in the presence of scarce data that we call example-based NER. Our train-free few-shot learning takes inspiration from question-answering identify spans new and unseen domain. In comparison with current state-of-the-art, proposed method performs significantly better, especially when using low number support examples.
Machine Learning is transitioning from an art and science into a technology available to every developer. In the near future, application on platform will incorporate trained models encode data-based decisions that would be impossible for developers author. This presents significant engineering challenge, since currently data modeling are largely decoupled standard software development processes. separation makes incorporating machine learning capabilities inside applications unnecessarily...
Tone mapping operators (TMO) are functions that map high dynamic range (HDR) images to a standard (SDR), while aiming preserve the perceptual cues of scene govern its visual quality. Despite increasing number studies on quality assessment tone mapped images, current subjective datasets have relatively small numbers and opinions. Moreover, existing challenges in transferring laboratory experiments crowdsourcing platforms put barrier for collecting large-scale through crowdsourcing. In this...
A Tone Mapping Operator (TMO) aims at reproducing the visual perception of a scene with high dynamic range (HDR) on low (LDR) media. TMOs have primarily aimed to preserve global by employing model human system (HVS), analysing perceptual attributes each pixel and adjusting exposure level.Preserving semantic perception, also an essential step for HDR rendering, has never been in explicit focus. We argue that explicitly introducing information create 'content semantic'-aware TMO potential...
Infrared (IR) imaging offers advantages in several fields due to its unique ability of capturing content extreme light conditions. However, the demanding hardware requirements high-resolution IR sensors limit widespread application. As an alternative, visible can be used synthesize images but this causes a loss fidelity image details and introduces inconsistencies lack contextual awareness scene. This stems from combination using with standard dynamic range, especially under lighting, result...
Tone mapping operators (TMO) are functions which map high dynamic range (HDR) images to limited media while aiming preserve the perceptual cues of scene that govern its aesthetic quality. Evaluating quality TMOs is non-trivial due subjectivity preference involved. Traditionally, TMO has been evaluated via subjective experiments in a controlled laboratory environment. However, last decade brought surge popularity crowdsourcing as an alternative methodology conduct experiments. uncontrolled...
Active extravasation into the upper aerodigestive tract is a dramatic and potentially life-threatening complication in patients with head neck cancers. It prompts presentation to emergency room subsequent urgent imaging identify source of hemorrhage. Imaging these may be complicated by treatment-altered anatomy, posing challenge radiologist who needs rapidly presence active hemorrhage potential vessel. This retrospective review summarizes clinical findings 6 oropharyngeal oral cavity...
This abstract introduces an innovative system for automatic payments using face recognition, eliminating the need traditional payment cards. In this proposed system, users authenticate transactions by presenting their faces to recognition prioritizing Face ID its effectiveness in identification. The technology relies on advanced biometric techniques, including OpenCV image processing, Haar Cascade Classifier detection, and Local Binary Pattern facial recognition. Upon successful...
The range of real-world scene luminance is larger than the capture capability many digital camera sensors which leads to details being lost in captured images, most typically bright regions. Inverse tone mapping attempts boost these Standard Dynamic Range (SDR) images back High (HDR) by creating a that linearizes well exposed values from SDR image, and provides clipped content. However, cases, regions cannot be recovered or estimated. In this paper, we present novel inverse approach for HDR...
A Tone Mapping Operator (TMO) is required to render images with a High Dynamic Range (HDR) on media limited dynamic capabilities.TMOs compress the range aim of preserving visually perceptual cues scene.Previous literature has established benefits TMOs being semantic aware, understanding content in scene preserve better.Expert photographers analyze and contextual information decide tonal transformations or local luminance adjustments.This process can be considered manual analogy tone...
In this paper, we introduce a novel approach for optimizing the environmental footprint of Business Processes by using techniques from area Compliance Checking along with set ecological weakness patterns that were identified during prior research. Through our approach, are able to identify weaknesses in process models can be revised come up modified, more eco-efficient process. work employ existing and divide them into clusters or genres based on their semantical meaning. Consequently,...
Tone mapping operators (TMO) are pivotal in rendering High Dynamic Range (HDR) content on limited dynamic range media.Analysing the quality of tone mapped images depends several objective factors and a combination subjective like aesthetics, fidelity etc. Objective Image assessment (IQA) metrics often used to evaluate TMO but they do not always reflect ground truth.A robust alternative IQA is assessment.Although, experiments provide accurate results, can be time-consuming expensive...
A Tone Mapping Operator (TMO) is required to render images with a High Dynamic Range (HDR) on media limited dynamic capabilities. TMOs compress the range aim of preserving visually perceptual cues scene. Previous literature has established benefits being semantic aware, understanding content in scene preserve better. Expert photographers analyze and contextual information decide tonal transformations or local luminance adjustments. This process can be considered manual analogy tone mapping....