Neural Network Image Scaling Using Spatial Errors
نویسندگان
چکیده
We propose a general method for gradient-based training of neural network (NN) models to scale multidimensional signal data. In the case of image data, the goal is to fit models that produce images of high perceptual quality, as opposed to simply a high peak signal to noise ratio (PSNR). There have been a number of perceptual image error measures proposed in the literature, the majority of which consider the behavior of the error surface in some local neighborhood of each pixel. By integrating such error measures into the NN learning framework, we may fit models that minimize the perceptual error, producing results that are more visually pleasing. We introduce a spatial error measure and discuss in detail the derivative computations necessary for backpropagation. The results are compared to neural networks trained with the standard sum of squared errors (SSE) function, as well as a state of the art scaling method.
منابع مشابه
An Automated MR Image Segmentation System Using Multi-layer Perceptron Neural Network
Background: Brain tissue segmentation for delineation of 3D anatomical structures from magnetic resonance (MR) images can be used for neuro-degenerative disorders, characterizing morphological differences between subjects based on volumetric analysis of gray matter (GM), white matter (WM) and cerebrospinal fluid (CSF), but only if the obtained segmentation results are correct. Due to image arti...
متن کاملVolumetric soil moisture estimation using Sentinel 1 and 2 satellite images
Surface soil moisture is an important variable that plays a crucial role in the management of water and soil resources. Estimating this parameter is one of the important applications of remote sensing. One of the remote sensing techniques for precise estimation of this parameter is data-driven models. In this study, volumetric soil moisture content was estimated using data-driven models, suppor...
متن کاملComparison of Artificial Neural Network Training Algorithms for Predicting the Weight of Kurdi Sheep using Image Processing
Extended Abstract Introduction and Objective: Due to weakness, the occurrence of unwanted errors, the impact of the environment and exposure to natural events, human always make mistakes in their diagnoses of the environment or different topics, so that different people 's perception of a single and unique event may be very different and be diverse. Nowadays, with the development of image proc...
متن کاملDesign, Development and Evaluation of an Orange Sorter Based on Machine Vision and Artificial Neural Network Techniques
ABSTRACT- The high production of orange fruit in Iran calls for quality sorting of this product as a requirement for entering global markets. This study was devoted to the development of an automatic fruit sorter based on size. The hardware consisted of two units. An image acquisition apparatus equipped with a camera, a robotic arm and controller circuits. The second unit consisted of a robotic...
متن کاملMotion detection by a moving observer using Kalman filter and neural network in soccer robot
In many autonomous mobile applications, robots must be capable of analyzing motion of moving objects in their environment. Duringmovement of robot the quality of images is affected by quakes of camera which cause high errors in image processing outputs. In thispaper, we propose a novel method to effectively overcome this problem using Neural Networks and Kalman Filtering theory. Thistechnique u...
متن کامل