Abstract In the last two decades, improvements in materials, sensors and machine learning technologies have led to a rapid extension of electronic nose (EN) related research topics with diverse applications. The food beverage industry, agriculture forestry, medicine health-care, indoor outdoor monitoring, military civilian security systems are leading fields which take great advantage from rapidity, stability, portability compactness ENs. Although EN technology provides numerous benefits,...

Assessing the quality of seafood both in retail and during packaging at production side must be carried out minutely order to avoid spoilage which causes severe human health problems also economic loss. Since illnesses decay presents distinct symptoms different species, primarily classification species is required. In this field, inadequacy current laborious slow traditional methods can overcome with systems based on machine learning image processing, present fast precise results. design...

Automatic classification of food freshness plays a significant role in the industry. Food spoilage detection from production to consumption stages needs be performed minutely. Traditional methods which detect are slow, laborious, subjective and time consuming. As result, fast accurate automatic need introduced industrial applications. This study comparatively analyses an image dataset containing samples three types fruits distinguish fresh those rotten. The proposed vision based framework...

High dynamic range (HDR) imaging enables to immortalize natural scenes similar the way that they are perceived by human observers. With regular low (LDR) capture/display devices, significant details may not be preserved in images due huge of scenes. To minimize information loss and produce high quality HDR-like for LDR screens, this study proposes an efficient multi-exposure fusion (MEF) approach with a simple yet effective weight extraction method relying on principal component analysis,...

In this paper, a human eye localization algorithm in images and video is presented for faces with frontal pose upright orientation. A given face region filtered by high-pass filter of wavelet transform. way, edges the are highlighted, caricature-like representation obtained. After analyzing horizontal projections profiles edge regions image, candidate points each detected. All then classified using support vector machine based classifier. Locations estimated according to most probable ones...

This paper describes two new intraimage prediction methods based on data dimensionality reduction methods: nonnegative matrix factorization (NMF) and locally linear embedding. These aim at approximating a block to be predicted in the image as combination of <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">k</i> -nearest neighbors determined known pixels causal neighborhood input block. Variable can seen parameter controlling some sort sparsity...

The paper presents a dictionary construction method for spatial texture prediction based on sparse approximations. Sparse approximations have been recently considered image using static dictionaries such as DCT or DFT dictionary. These approaches rely the assumption that is periodic, hence use of formed by pre-defined waveforms. However, in real images, there are more complex and non-periodic textures. main idea underlying proposed technique instead to consider locally adaptive dictionary,...

We describe a self-content single-image super-resolution algorithm based on multi-scale neighbor embeddings of small image patches. Given an input low-resolution patch, we gradually expand its size by relying local geometric similarities low- and high-resolution patch spaces under scaling factors. characterize the geometry with K-similar patches taken from exemplar set collect pairs appropriately rescaled versions. While ensuring images compatibility optimization K, satisfy smoothness...

Under unsuitable sales conditions, red meat containing rich amount of protein might receive a negative perception from consumers. Importantly, nutrients lose their effectiveness, while at the same time formation harmful microorganisms becomes threat to human health. The main purpose this study is keep quality open department service offered consumers in retail sector highest level, ensure sustainability resources and provide immediate economic precautions by reducing disposal due possible...

In this paper, a human face detection method in images and video is presented. After determining possible candidate regions using color information, each region filtered by high-pass filter of wavelet transform. way, edges the are highlighted, caricature-like representation obtained. Horizontal, vertical filter-like projections used as feature signals dynamic programming (DP) support vector machine (SVM) based classifiers. It turns out that classifier provides better rates compared to our...

The paper first describes an examplar-based image inpainting algorithm using a locally linear neighbor embedding technique with low-dimensional neighborhood representation (LLE-LDNR). searches the K nearest neighbors ( ) of input patch to be filled-in and linearly combine them LLE-LDNR synthesize missing pixels. Linear regression is then introduced for improving K-NN search. performance enhanced search method assessed two applications: loss concealment object removal.

We describe a self-content single image super-resolution algorithm based on multi-scale neighbor embeddings of patches. make use the recurrence property similar patches across different scales an image. Inspired by manifold learning approaches, we first characterize local geometry given low-resolution patch reconstructing it from taken down-scaled versions input then hallucinate high-resolution relying geometric similarities low- and spaces. enforce compatibility through overlapping,...

The template matching algorithm is a simple extension to exemplar-based texture synthesis. Average of predictors or non-local means based approaches can be seen as heuristic extensions matching. These methods which linearly combine several patches have been shown more robust in synthesis and give better results when compared However, they do not search minimize an approximation error on the known pixel values template. They are rather for calculating linear weighting coefficients. This paper...

We propose an exemplar-based super-resolution algorithm based on sparsity constrained neighbor-embeddings of local image patches. extract exemplar patch pairs from as little the given low-resolution image, and we rely geometric similarities low-and high-resolution spaces. While sparsely coding geometry with a greedy selection method, refine our solution by iteratively updating obtained image. finally apply adaptive back-projection to ensure global consistency. Our experimental results...

This paper presents a novel dictionary learning method which, because of its simplicity and the limited number training samples it requires, can be used for online dictionaries spatial texture prediction. The proposed has first been described to address problem intra image prediction based on signal expansion overcomplete dictionaries. It then evaluated in complete codec. experimental results obtained show significant improvement terms quality predicted compared H.264/AVC Significant...

Melanoma which occurs with non-healing DNA degradation in melanocyte cells, is the most deadly type of skin cancers. Importantly, it can be identified for a treatment before spreads to other tissues, i.e., early diagnosis. To identify, specialist visually inspects whether suspected lesion melanoma or not. However, due different education and experience levels specialists as result patient not being facility that specialized this area, problem "subjectivity" arises, good visual investigation...

The paper studies several non-negative matrix factorization methods with nearest neighbors constrained dictionaries for image prediction. considered include the multiplicative update algorithm, projected gradient as well graph-regularized NMF solution which aims at taking into account geometrical structure of input data. Intra prediction problem based on these solutions amounts to a neighbor embedding problem. Both and rate-distortion performances are then given in comparison other like...