The Semantic Pixel

نویسندگان

  • Frederico T. Fonseca
  • Clodoveu A. Davis
  • Gilberto Câmara
چکیده

Usually, images can be seen as sets of pixels or as fields over a reference space. While the former view allows image processing to function using pixel manipulation algorithms, the second one is closer to a wider understanding of what people perceive in an image. The pixel aspect is much closer to the measurement, to observations, while fields are closer to the semantic aspect, to the interpretation of the observations. This paper discusses some semantic challenges related to integration of image data from various sources, considering both views. Such integration is necessary, considering that soon a new generation of remote sensing satellites based on free and open data policies is expected to become operational, so researchers will have access to more data than they can handle with current techniques. We propose the integration of images from multiple sensors starting from a common point, which we call the Semantic Pixel. It will enable scientists to have access to large sets of satellite images and their metadata, regardless of source or format. The Semantic Pixel will also enable access to ancillary data, which is essential for advanced temporal analysis of forest cover dynamics, including major sets of natural resource data, such as vegetation, soil and geology maps. Other data encoded as fields, such as digital elevation models, relevant climatic variable maps, political maps and associated census data, can also fit this model. Resumo. Imagens podem ser vistas como um conjunto de pixels ou como campos em um espaço de referencia. Enquanto os pixels permitem que algoritmos de processamento de imagens possam funcionar, os campos estão mais próximos do que as pessoas entendam o que seja o significado de uma imagem. A visão de pixels está muito mais próximo das medidas, das observações, enquanto os campos estão mais próximos da semântica, da interpretação das observações. Este artigo usa estas duas visões para discutir os desafios relativos a semântica na questão das integração de imagens de provenientes de fontes diversas. Esta integração é necessária já que brevemente novos satélites com politicas de dados abertos devem estar disponíveis o que levará os cientistas a terem mais dados do que eles possam efetivamente usar. Aqui nós apresentamos uma proposta de integração de imagens de fontes diversas usando como plataforma inicial um ponto comum, que chamamos o Pixel Semântico. Desta maneira os cientistas teriam acesso a dados e metadados de imagens independente da fonte ou formato. O Pixel Semântico também possibilitaria o acesso a dados históricos que são importantes para análises da dinâmica da vida das florestas tropicais.

برای دانلود رایگان متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Semantic image understanding : from pixel to word

The aim of semantic image understanding is to reveal the semantic meaning behind the image pixel. We categorize semantic image understanding into two broad categories: pixel-level and image-level semantic image understanding. While pixel-level image understanding aims to obtain the semantic meaning of each pixel, image-level understanding aims to obtain the semantic meaning of the whole image, ...

متن کامل

Deep Context Convolutional Neural Networks for Semantic Segmentation

Recent years have witnessed the great progress for semantic segmentation using deep convolutional neural networks (DCNNs). This paper presents a novel fully convolutional network for semantic segmentation using multi-scale contextual convolutional features. Since objects in natural images tend to be with various scales and aspect ratios, capturing the rich contextual information is very critica...

متن کامل

Transferable Semi-supervised Semantic Segmentation

The performance of deep learning based semantic segmentation models heavily depends on sufficient data with careful annotations. However, even the largest public datasets only provide samples with pixel-level annotations for rather limited semantic categories. Such data scarcity critically limits scalability and applicability of semantic segmentation models in real applications. In this paper, ...

متن کامل

Pixel-Level Encoding and Depth Layering for Instance-Level Semantic Labeling

Recent approaches for instance-aware semantic labeling have augmented convolutional neural networks (CNNs) with complex multitask architectures or computationally expensive graphical models. We present a method that leverages a fully convolutional network (FCN) to predict semantic labels, depth and an instance-based encoding using each pixel’s direction towards its corresponding instance center...

متن کامل

On the Pixel Expansion of Hypergraph Access Structures in Visual Cryptography Schemes

In a visual cryptography scheme, a secret image is encoded into n shares, in the form of transparencies. The shares are then distributed to n participants. Qualified subsets of participants can recover the secret image by superimposing their transparencies, but non-qualified subsets of participants have no information about the secret image. Pixel expansion, which represents the number of subpi...

متن کامل

Pixel and semantic capabilities from an image-object based document representation

This paper reports on novel and traditional pixel and semantic operations using a recently standardized document representation called JPM. The JPM representation uses compressed pixel arrays for all visible elements on a page. Separate data containers called boxes provide the layout and additional semantic information. JPM and related image-based document representation standards were designed...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 2014