ESTG - Mestrado em Engenharia Eletrotécnica - Telecomunicações
Permanent URI for this collection
Browse
Browsing ESTG - Mestrado em Engenharia Eletrotécnica - Telecomunicações by advisor "Faria, Sérgio Manuel Maciel de"
Now showing 1 - 8 of 8
Results Per Page
Sort Options
- Disparity compensation using geometric transformsPublication . Monteiro, Ricardo Jorge Santos; Rodrigues, Nuno Miguel Morais; Faria, Sérgio Manuel Maciel deThis dissertation describes the research and development of some techniques to enhance the disparity compensation in 3D video compression algorithms. Disparity compensation is usually performed using a block matching technique between views, disregarding the various levels of disparity present for objects at different depths in the scene. An alternative coding scheme is proposed, taking advantage of the cameras setup information and the object’s depth in the scene, to compensate more complex spatial distortions, being able to improve disparity compensation even with convergent cameras. In order to perform a more accurate disparity compensation, the reference picture list is enriched with additional geometrically transformed images, for the most relevant object’s levels of depth in the scene, resulting from projections of one view to another. This scheme can be implemented in any state-of-the-art video codec, as H.264/AVC or HEVC, in order to improve the disparity matching accuracy between views. Experimental results, using MV-HEVC extension, show the efficiency of the proposed method for coding stereo video, presenting bitrate savings up to 2.87%, for convergent camera sequences, and 1.52% for parallel camera sequences. Also a method to choose the geometrically transformed inter view reference pictures was developed, in order to reduce unnecessary overhead for unused reference pictures. By selecting and adding to the reference picture list, only the most useful pictures, all results improved, presenting bitrate savings up to 3.06% for convergent camera sequences, and 2% for parallel camera sequences.
- Improving minimum rate predictors algorithm for compression of volumetric medical imagesPublication . Santos, João Miguel Pereira da Silva; Faria, Sérgio Manuel Maciel de; Rodrigues, Nuno Miguel MoraisMedical imaging technologies are experiencing a growth in terms of usage and image resolution, namely in diagnostics systems that require a large set of images, like CT or MRI. Furthermore, legal restrictions impose that these scans must be archived for several years. These facts led to the increase of storage costs in medical image databases and institutions. Thus, a demand for more efficient compression tools, used for archiving and communication, is arising. Currently, the DICOM standard, that makes recommendations for medical communications and imaging compression, recommends lossless encoders such as JPEG, RLE, JPEG-LS and JPEG2000. However, none of these encoders include inter-slice prediction in their algorithms. This dissertation presents the research work on medical image compression, using the MRP encoder. MRP is one of the most efficient lossless image compression algorithm. Several processing techniques are proposed to adapt the input medical images to the encoder characteristics. Two of these techniques, namely changing the alignment of slices for compression and a pixel-wise difference predictor, increased the compression efficiency of MRP, by up to 27.9%. Inter-slice prediction support was also added to MRP, using uni and bi-directional techniques. Also, the pixel-wise difference predictor was added to the algorithm. Overall, the compression efficiency of MRP was improved by 46.1%. Thus, these techniques allow for compression ratio savings of 57.1%, compared to DICOM encoders, and 33.2%, compared to HEVC RExt Random Access. This makes MRP the most efficient of the encoders under study.
- LEARNING-BASED IMAGE COMPRESSION USING MULTIPLE AUTOENCODERSPublication . António, Rúben Duarte; Assunção, Pedro António Amado; Faria, Sérgio Manuel Maciel de; Távora, Luís Miguel de Oliveira Pegado de Noronha eAdvanced video applications in smart environments (e.g., smart cities) bring different challenges associated with increasingly intelligent systems and demanding requirements in emerging fields such as urban surveillance, computer vision in industry, medicine and others. As a consequence, a huge amount of visual data is captured to be analyzed by task-algorithm driven machines. Due to the large amount of data generated, problems may occur at the data management level, and to overcome this problem it is necessary to implement efficient compression methods to reduce the amount of stored resources. This thesis presents the research work on image compression methods using deep learning algorithms analyzing the properties of different algorithms, because recently these have shown good results in image compression. It is also explained the convolutional neural networks and presented a state-of-the-art of autoencoders. Two compression approaches using autoencoders were studied, implemented and tested, namely an object-oriented compression scheme, and algorithms oriented to high resolution images (UHD and 360º images). In the first approach, a video surveillance scenario considering objects such as people, cars, faces, bicycles and motorbikes was regarded, and a compression method using autoencoders was developed with the purpose of the decoded images being delivered for machine vision processing. In this approach the performance was measured analysing the traditional image quality metrics and the accuracy of task driven by machine using decoded images. In the second approach, several high resolution images were considered adapting the method used in the previous approach considering properties of the image, like variance, gradients or PCA of the features, instead of the content that the image represents. Regarding the first approach, in comparison with the Versatile Video Coding (VVC) standard, the proposed approach achieves significantly better coding efficiency, e.g., up to 46.7% BD-rate reduction. The accuracy of the machine vision tasks is also significantly higher when performed over visual objects compressed with the proposed scheme in comparison with the same tasks performed over the same visual objects compressed with the VVC. These results demonstrate that the learningbased approach proposed is a more efficient solution for compression of visual objects than standard encoding. Considering the second approach although it is possible to obtain better results than VVC on the test subsets, the presented approach only presents significant gains considering 360º images.
- LOSSY COMPRESSION OF BIOMEDICAL IMAGES FOR COMPUTER VISION ANALYSISPublication . Paulo, Edgar da Silva; Faria, Sérgio Manuel Maciel de; Távora, Luís Miguel de Oliveira Pegado de Noronha e; Thomaz, Lucas ArrabalThe exponential increase in medical and biomedical data acquisition is compelled by technological advances, namely in the imaging field. However, this exponential growth brings with it challenges in terms of processing capacity, transmission, and data storage. In response to this growing demand, increasingly efficient solutions have emerged, especially through computer vision for automatic image analysis and compression algorithms. This dissertation aims, on the one hand, to evaluate the performance of computer vision systems on previously compressed biomedical images. On the other hand, it increases the useful range of image variations, almost lossless and lossy, decreasing the impact of the change added by this method on the performance of computer vision algorithms in biomedical image analysis. In this sense, YOLO and Detectron2 are employed to evaluate the impact of coding distortion on their ability to detect mitochondria in electron microscopy images. The results of this study reveal that although the distortion introduced by compression affects their detection performance, it is negligible at lower compression ratios. Furthermore, two proposals are presented to improve the useful compression ratio, keeping the images characteristics that allow to perform the automatic detection of mitochondria. On the one hand, it is demonstrated that the proposed training methodology, which incorporates compressed versions of the original data during training, mitigates the impact of distortion on the performance of computer vision algorithms; on the other hand, allocating higher quality levels to regions of interest, compared to background elements, helps to sustain high performance at compression rates where computer vision algorithms typically start to lose effectiveness. These approaches allow the extension of the compression range with little impact on detection performance, thus contributing to the improvement of data processing, storage, and transmission in biomedical applications.
- M U LT I - S E N S O R S M A RT PA C K A G E T R A C K I N G D E V I C E F O R L O G I S T I C SPublication . Susano, Rúben Rafael Pinto; Fernandes, Telmo Rui Carvalhinho Cunha; Faria, Sérgio Manuel Maciel de; Ventura, Paulo Jorge da CruzA embalagem comum tem recentemente vindo a demonstrar insuficiências na aptidão para satisfazer as expectativas do público em relação a produtos frescos, de alta qualidade e seguros. Com isto, a indústria obteve motivação para inovar e procurar novas soluções para corrigir esta lacuna. Através de estudos nesta área, a Embalagem Inovadora surgiu como uma solução viável. Este conceito foca-se principalmente em melhorar embalagens, através da integração de tecnologias de diversas áreas. O seu principal desafio reside em melhorar as embalagens, enquanto preserva ou melhora o fluxo de trabalho da cadeia de abastecimento e logística. Esta dissertação apresenta uma solução de Embalagem Inovadora, baseada em eletrónica, para monitorizar e rastrear embalagens numa cadeia de distribuição. O estudo aborda tanto o projeto como o desenvolvimento do hardware e do firmware, apresentando também os testes de desempenho realizados com vista à validação da solução desenvolvida. A solução desenvolvida consiste num sistema capaz de monitorizar temperatura, humidade, pressão atmosférica, impactos e localização das embalagens. O sistema está dividido em Package Tags e Gateways. As Package Tags encontram-se situadas dentro da embalagem e mantêm o registo de todos os valores mencionados, excluindo a localização. Os Gateways funcionam como pontes entre várias Package Tags e o servidor remoto, servindo igualmente para determinar a localização a ser associada à Package Tag. A comunicação entre as Package Tags e os Gateways é realizada utilizando o recurso PAwR do protocolo BLE. Embora este recurso tenha sido originalmente desenvolvido para etiquetas de preço ESL em supermercados, foi adaptado, durante este trabalho, para a sua aplicação em logística dentro deste sistema.
- ROI -BASED CODING OF BIOMEDICAL IMAGES FOR MACHINE ANALYSISPublication . Nicolau, Daniel Filipe da Silva; Faria, Sérgio Manuel Maciel de; Távora, Luís Miguel de Oliveira Pegado de Noronha e; Thomaz, Lucas ArrabalThe increasing volume of data acquired and generated daily in the healthcare sector, driven by technological advancements, brings significant benefits to patient diagnosis and research. However, this growth also presents considerable challenges in the analysis and processing of such data. To address these difficulties, computer vision algorithms have emerged as powerful tools, capable of automating repetitive and time-consuming tasks, enabling faster and more accurate processing. At the same time, the growing volume of data places pressure on storage and transmission capabilities, demanding efficient compression methods to minimise its size. In the literature, various approaches are found, primarily divided into two categories: lossy and lossless compression. While lossless methods ensure data integrity, they do not achieve compression rates as high as lossy algorithms. The latter, despite significantly reducing file sizes, introduces distortions that may compromise image quality, affecting the accuracy of automated systems. This dissertation focuses on two main challenges: first, evaluating the impact of image compression on the performance of biomedical computer vision systems, and second, improving compression efficiency without compromising the accuracy of these algorithms. To this end, detection and segmentation models, such as YOLOv8 and SAM, were used to analyse the effect of distortion caused by encoding on the localisation and segmentation of mitochondria in two datasets of electron microscopy images. To enhance model performance at higher compression levels, two methodologies were implemented. The first focuses on domain adaptation, fine-tuning the models to recognise and compensate for distortions introduced by compression, specifically in HEVC/H.265 and VVC/H.266 encoders. The second approach proposes contentaware encoder adaptation, allowing the assignment of different quality levels to selected regions of interest. This method aims to reduce storage and bandwidth requirements without significantly compromising the performance of deep learningbased models. Experimental results demonstrate that region-of-interest-based encoding strategies effectively reduce compression rates while maintaining model accuracy. In particular, the proposed methodologies allowed to achieve an average performance improvement of up to 23.70% for the same bpp range and a data size reduction of up to 74.96%. Additionally, a Pareto-based optimisation algorithm was proposed to determine the most suitable encoding configurations for different standards and models, ensuring a balance between compression efficiency and object detection performance.
- S I S T E M A D E M O N I T O R I Z A Ç Ã O AVA N Ç A D O PA R A A Q U A P O N I APublication . Pereira, Fernando de Jesus Santos; Faria, Sérgio Manuel Maciel de; Vieira, Judite dos Santos; Fernandes, Telmo Rui Carvalhinho CunhaA monitorização da qualidade da água em sistemas de aquaponia é essencial para garantir o equilíbrio entre os organismos aquáticos e as plantas nele inserido, permitindo a deteção de variações nos parâmetros essenciais ao bom funcionamento do sistema. Neste contexto, esta dissertação apresenta o desenvolvimento e validação de dois sistemas complementares de monitorização, um para a qualidade da água na estufa e outro para a água fornecida à estufa, ambos recorrendo a sensores e comunicação sem fios para permitir uma análise contínua e remota dos parâmetros monitorizados. O primeiro sistema, instalado na estufa, integra sensores para a medição da temperatura da água, pH, condutividade elétrica (CE), sólidos dissolvidos totais (TDS) e potencial de oxidação-redução (ORP), garantindo a monitorização das condições da água utilizada na produção. Além disso, foi implementado um sistema de monitorização visual, baseado em ESP32-CAM, para acompanhar o crescimento das plantas e detetar a presença de insetos. O segundo sistema, responsável pela monitorização da água antes da sua entrada na estufa, apresenta uma arquitetura semelhante, mas foi projetado para operar em ambientes externos, exigindo uma caixa à prova de água e comunicação via NB-IoT para transmissão remota dos dados. Os sensores utilizados foram submetidos a um processo de calibração regular, assegurando medições fiáveis e dentro dos valores recomendados para sistemas de aquaponia. A análise dos dados recolhidos demonstrou que os sistemas apresentam um desempenho estável ao longo do tempo, permitindo a deteção de variações nos parâmetros indicadores da qualidade da água e facilitando a tomada de decisões para a manutenção do equilíbrio do sistema. A integração da monitorização vi sual mostrou-se uma ferramenta complementar importante, permitindo validar as condições do crescimento das plantas e a presença de pragas de forma remota. Os resultados confirmam a eficácia dos sistemas desenvolvidos, reduzindo a necessidade de intervenções manuais e proporcionando uma solução eficiente e modular para a monitorização de ambientes aquapónicos.
- Sistema de Comunicação em Ambiente Aquático para Medição do Desempenho Físico em AtletasPublication . Santos, Miguel Oliveira; Faria, Sérgio Manuel Maciel de; Fernandes, Telmo Rui Carvalhinho Cunha