- Image and Video Quality Assessment
- Video Coding and Compression Technologies
- Advanced Image Processing Techniques
- Multimedia Communication and Technology
- Advanced Data Compression Techniques
- Advanced Photonic Communication Systems
- Advanced Image Fusion Techniques
- Advanced Vision and Imaging
- Caching and Content Delivery
- Image Retrieval and Classification Techniques
- VLSI and Analog Circuit Testing
- Machine Learning and Data Classification
- Experimental Learning in Engineering
- Advanced Bandit Algorithms Research
- Data Stream Mining Techniques
- Analog and Mixed-Signal Circuit Design
CEA LIST
2022-2024
Commissariat à l'Énergie Atomique et aux Énergies Alternatives
2022-2024
Université Paris-Saclay
2022-2024
Fraunhofer Institute for Telecommunications, Heinrich Hertz Institute
2023
University of Klagenfurt
2023
The rise in video streaming applications has increased the demand for quality assessment (VQA). In 2016, Netflix introduced Video Multi-Method Assessment Fusion (VMAF), a full reference VQA metric that strongly correlates with perceptual quality, but its computation is time-intensive. We propose Discrete Cosine Transform (DCT)-energy-based texture information fusion (VQ-TIF) model determines visual of reconstructed compared to original video. VQ-TIF extracts Structural Similarity (SSIM) and...
Adaptive live video streaming applications utilize a predefined collection of bitrate-resolution pairs, known as <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">bitrate ladder</i> , for simplicity and efficiency, eliminating the need additional run-time to determine optimal pairs during session. These do not incorporate two-pass encoding methods due increased latency. However, an optimized bitrate ladder could result in lower storage delivery...
In recent years, video streaming applications have proliferated the demand for Video Quality Assessment VQA). Reduced reference quality assessment (RR-VQA) is a category of VQA where certain features (e.g., texture, edges) original are provided assessment. It popular research area various such as social media, online games, and streaming. This paper introduces reduced Transcoding Prediction Model (TQPM) to determine visual score possibly transcoded in multiple stages. The predicted using...
Traditional per-title encoding schemes aim to optimize resolutions deliver the highest perceptual quality for each representation. However, keeping time within an acceptable threshold a smooth user experience is important reduce carbon footprint and energy consumption on servers in video streaming applications. Toward this realization, we introduce latency-aware dynamic resolution scheme (LADRE) adaptive LADRE determines target bitrate by utilizing random forest-based prediction model every...
In HTTP adaptive live streaming applications, video segments are encoded at a fixed set of bitrate-resolution pairs known as bitrate ladder. Live encoders use the fastest available encoding configuration, referred to preset, ensure minimum possible latency in encoding. However, an optimized preset and number CPU threads for each instance may result (i) increased quality (ii) efficient utilization while For low encoders, speed is expected be more than or equal framerate. To this light, paper...
Versatile Video Coding (VVC) allows for large compression efficiency gains over its predecessor, High Efficiency (HEVC). The added comes at the cost of increased runtime complexity, especially encoding. It is thus highly relevant to explore all available reduction options. This paper proposes a novel first pass two-pass rate control in all-intra configuration, using low-complexity video analysis and Random Forest (RF)-based machine learning model derive data required driving second pass....
Traditional per-title encoding schemes aim to optimize resolutions deliver the highest perceptual quality for each representation. XPSNR is observed correlate better with subjective of VVC-coded bitstreams. Towards this realization, we predict average bitstreams using spatiotemporal complexity features video and target configuration an XGBoost-based model. Based on predicted scores, introduce a Quality-A ware Dynamic Resolution Adaptation (QADRA) framework adaptive streaming applications,...
Adaptive live video streaming applications use a fixed predefined configuration for the bitrate ladder with constant framerate and encoding presets in session. However, selecting optimized framerates every representation can enhance perceptual quality, improve computational resource allocation, thus, energy efficiency. In particular, low low-bitrate representations reduce compression artifacts decrease consumption. addition, an preset may lead to improved To this light, paper proposes...
With the emergence of multiple modern video codecs, streaming service providers are forced to encode, store, and transmit bitrate ladders codecs separately, consequently suffering from additional energy costs for encoding, storage, transmission. To tackle this issue, we introduce an online energy-efficient Multi-Codec Bitrate ladder Estimation scheme (MCBE) adaptive applications. In MCBE, quality representations within new-generation (e.g., High Efficiency Video Coding (HEVC), Alliance Open...
<p>Adaptive live video streaming applications utilize a predefined collection of bitrate-resolution pairs, known as bitrate ladder, for simplicity and efficiency, eliminating the need additional run-time to determine optimal pairs each video. These do not incorporate two-pass encoding methods due increased latency. However, an optimized ladder could result in lower storage delivery costs improved Quality Experience (QoE). This paper presents Just Noticeable Difference (JND)-aware...
Conventional per-title encoding schemes strive to optimize resolutions deliver the utmost perceptual quality for each bitrate ladder representation. Nevertheless, maintaining time within an acceptable threshold is equally imperative in online streaming applications. Furthermore, modern client devices are equipped with capability fast deep-learning-based video super-resolution (VSR) techniques, enhancing of decoded bitstream. This suggests that opting lower representations during process can...
Conventional per-title encoding schemes strive to optimize resolutions deliver the utmost perceptual quality for each bitrate ladder representation. Nevertheless, maintaining time within an acceptable threshold is equally imperative in online streaming applications. Further-more, modern client devices are equipped with capability fast deep-learning-based video super-resolution (VSR) techniques, enhancing of decoded bitstream. This suggests that opting lower representations during process can...
<p>Adaptive live video streaming applications utilize a predefined collection of bitrate-resolution pairs, known as bitrate ladder, for simplicity and efficiency, eliminating the need additional run-time to determine optimal pairs each video. These do not incorporate two-pass encoding methods due increased latency. However, an optimized ladder could result in lower storage delivery costs improved Quality Experience (QoE). This paper presents Just Noticeable Difference (JND)-aware...
The rise in video streaming applications has increased the demand for quality assessment (VQA). In 2016, Netflix introduced Video Multi-Method Assessment Fusion (VMAF), a full reference VQA metric that strongly correlates with perceptual quality, but its computation is time-intensive. We propose Discrete Cosine Transform (DCT)-energy-based texture information fusion (VQ-TIF) model determines visual of reconstructed compared to original video. VQ-TIF extracts Structural Similarity (SSIM) and...
Versatile Video Coding (VVC) allows for large compression efficiency gains over its predecessor, High Efficiency (HEVC). The added comes at the cost of increased runtime complexity, especially encoding. It is thus highly relevant to explore all available reduction options. This paper proposes a novel first pass two-pass rate control in all-intra configuration, using low-complexity video analysis and Random Forest (RF)-based machine learning model derive data required driving second pass....
With the emergence of multiple modern video codecs, streaming service providers are forced to encode, store, and transmit bitrate ladders codecs separately, consequently suffering from additional energy costs for encoding, storage, transmission. To tackle this issue, we introduce an online energy-efficient Multi-Codec Bitrate ladder Estimation scheme (MCBE) adaptive applications. In MCBE, quality representations within new-generation (e.g., High Efficiency Video Coding (HEVC), Alliance Open...
This paper proposes Stream-based Active Learning method using Adaptive uncertainty and Diversity Thresholds (SALAT), which aims to reduce the amount of data required for labeling in stream-based settings while accounting concept drift. The proposed uses adaptive diversity thresholds, are monitored by a sliding window selection history ensure that is neither too conservative nor aggressive. effectiveness SALAT validated on MNIST FashionMNIST datasets, where it achieves similar performance...