A5026966525- Biometric Identification and Security
- Face recognition and analysis
- Retinal Imaging and Analysis
- Vehicle License Plate Recognition
- User Authentication and Security Systems
- Ocular Disorders and Treatments
- Handwritten Text Recognition Techniques
- Smart Parking Systems Research
- Advanced Neural Network Applications
- Plant tissue culture and regeneration
- Forensic and Genetic Research
- Traffic control and management
- Advanced Image and Video Retrieval Techniques
- Autonomous Vehicle Technology and Safety
- Forensic Fingerprint Detection Methods
- Ocular Diseases and Behçet’s Syndrome
- Smart Agriculture and AI
- Berry genetics and cultivation research
- Reconstructive Facial Surgery Techniques
- Innovations in Aquaponics and Hydroponics Systems
- Cutaneous Melanoma Detection and Management
- Greenhouse Technology and Climate Control
- Retinal and Optic Conditions
- Image and Object Detection Techniques
- Mobile and Web Applications
Universidade Federal do Paraná
2017-2023
Universidade Estadual de Ponta Grossa
2015-2016
Automatic License Plate Recognition (ALPR) has been a frequent topic of research due to many practical applications. However, the current solutions are still not robust in real-world situations, commonly depending on constraints. This paper presents and efficient ALPR system based state-of-the-art YOLO object detector. The Convolutional Neural Networks (CNNs) trained finetuned for each stage so that they under different conditions (e.g., variations camera, lighting, background). Specially...
Abstract This paper presents an efficient and layout‐independent Automatic License Plate Recognition (ALPR) system based on the state‐of‐the‐art you only look once (YOLO) object detector that contains a unified approach for license plate (LP) detection layout classification to improve recognition results using post‐processing rules. The is conceived by evaluating optimizing different models, aiming at achieving best speed/accuracy trade‐off each stage. networks are trained images from...
Existing approaches for image-based Automatic Meter Reading (AMR) have been evaluated on images captured in well-controlled scenarios. However, real-world meter reading presents unconstrained scenarios that are way more challenging due to dirt, various lighting conditions, scale variations, in-plane and out-of-plane rotations, among other factors. In this work, we present an end-to-end approach AMR focusing Our main contribution is the insertion of a new stage pipeline, called corner...
The iris is considered as the biometric trait with highest unique probability. location an important task for biometrics systems, affecting directly results obtained in specific applications such recognition, spoofing and contact lenses detection, among others. This work defines problem delimitation of smallest squared window that encompasses region. In order to build a benchmark we annotate (iris bounding boxes) four databases from different make them publicly available community. Besides...
There are several biometric-based systems which rely on a single biometric modality, most of them focus face, iris or fingerprint. Despite the good accuracies obtained with modalities, these more susceptible to attacks, i.e, spoofing and noises all kinds, especially in non-cooperative (in-the-wild) environments. Since environments becoming common, new approaches involving multi-modal biometrics have received attention. One challenge multimodal is how integrate data from different modalities....
Bias and fairness of biometric algorithms have been key topics research in recent years, mainly due to the societal, legal ethical implications potentially unfair decisions made by automated decision-making models. A considerable amount work has done on this topic across different modalities, aiming at better understanding main sources algorithmic bias or devising mitigation measures. In work, we contribute these efforts present first study investigating sclera segmentation Although...
One of the major challenges in ocular biometrics is cross-spectral scenario, i.e., how to match images acquired different wavelengths (typically visible (VIS) against near-infrared (NIR)). This article designs and extensively evaluates verification methods, for both closed open-world settings, using well known deep learning representations based on iris periocular regions. Using as inputs bounding boxes non-normalized iris/periocular regions, we fine-tune Convolutional Neural Network(CNN)...
The paper presents a summary of the 2020 Sclera Segmentation Benchmarking Competition (SSBC), 7th in series group benchmarking efforts centred around problem sclera segmentation. Different from previous editions, goal SSBC was to evaluate performance sclera-segmentation models on images captured with mobile devices. competition used as platform assess sensitivity existing i) differences devices for image capture and ii) changes ambient acquisition conditions. 26 research groups registered...
Recently, ocular biometrics in unconstrained environments using images obtained at visible wavelength have gained the researchers' attention, especially with captured by mobile devices. Periocular recognition has been demonstrated to be an alternative when iris trait is not available due occlusions or low image resolution. However, periocular does high uniqueness presented trait. Thus, use of datasets containing many subjects essential assess biometric systems' capacity extract...
The use of iris as a biometric trait is widely used because its high level distinction and uniqueness. Nowadays, one the major research challenges relies on recognition images obtained in visible spectrum under unconstrained environments. In this scenario, acquired are affected by capture distance, rotation, blur, motion low contrast specular reflection, creating noises that disturb systems. Besides delineating region, usually preprocessing techniques such normalization segmentation noisy...
ABSTRACT Micropropagation of small fruits such as blackberry has been employed due to the need obtain plants with high phytosanitary quality. Bioreactor technology used improve efficiency in seedling production. Thus, objective this work was evaluate best culture medium volume and sucrose concentration for micropropagation a temporary immersion bioreactor. In vitro shoots were segmented containing two buds an internode (1.0 cm) placed into MS supplemented inositol (0.1 g L-1), BAP (1 mg L-1)...
Ocular biometric systems working in unconstrained environments usually face the problem of small within-class compactness caused by multiple factors that jointly degrade quality obtained data. In this work, we propose an attribute normalization strategy based on deep learning generative frameworks, reduces variability samples used pairwise comparisons, without reducing their discriminability. The proposed method can be seen as a preprocessing step contributes for data regularization and...
In this work, we propose to detect the iris and periocular regions simultaneously using coarse annotations two well-known object detectors: YOLOv2 Faster R-CNN. We believe can be used in recognition systems based on regions, given much smaller engineering effort required manually annotate training images. made of (122K images from visible (VIS) spectrum 38K near-infrared (NIR) spectrum). The NIR databases were generated semi-automatically by first applying an segmentation CNN then performing...
The Artificial Neural Network (ANN) approach has had been applied in solutions for several problems such as classification, prediction and pattern recognition. Mostly, these belong to diversified areas of knowledge, not necessarily related Computer Science, e.g. agriculture, specifically precision recognition diseases plants, among others. Therefore, it is possible visualize the multidisciplinary feature that ANNs have, motivating development tools simulating more generic ANN, making its use...
In smart cities, it is common practice to define a maximum length of stay for given parking space increase the space's rotativity and discourage usage individual transportation solutions. However, automatically determining car dwell times from images faces challenges, such as collected low-resolution cameras, lighting variations, weather effects. this work, we propose method that combines two deep neural networks compute time each in lot. The proposed first defines status between occupied...