Search results for "artificial intelligence"
showing 10 items of 6122 documents
Eye movements when reading sentences with handwritten words.
2016
The examination of how we read handwritten words (i.e., the original form of writing) has typically been disregarded in the literature on reading. Previous research using word recognition tasks has shown that lexical effects (e.g., the word-frequency effect) are magnified when reading difficult handwritten words. To examine this issue in a more ecological scenario, we registered the participants’ eye movements when reading handwritten sentences that varied in the degree of legibility (i.e., sentences composed of words in easy vs. difficult handwritten style). For comparison purposes, we included a condition with printed sentences. Results showed a larger reading cost for sentences with dif…
Tetrachromatic color vision in goldfish: evidence from color mixture experiments
1992
Additive color mixture experiments were performed in the goldfish using a behavioral training technique in which the fish had to discriminate between two test fields.
Precomputed Real-Time Texture Synthesis with Markovian Generative Adversarial Networks
2016
This paper proposes Markovian Generative Adversarial Networks (MGANs), a method for training generative networks for efficient texture synthesis. While deep neural network approaches have recently demonstrated remarkable results in terms of synthesis quality, they still come at considerable computational costs (minutes of run-time for low-res images). Our paper addresses this efficiency issue. Instead of a numerical deconvolution in previous work, we precompute a feed-forward, strided convolutional network that captures the feature statistics of Markovian patches and is able to directly generate outputs of arbitrary dimensions. Such network can directly decode brown noise to realistic textu…
Leveraging Uncertainty Estimates to Improve Segmentation Performance in Cardiac MR
2021
In medical image segmentation, several studies have used Bayesian neural networks to segment and quantify the uncertainty of the images. These studies show that there might be an increased epistemic uncertainty in areas where there are semantically and visually challenging pixels. The uncertain areas of the image can be of a great interest as they can possibly indicate the regions of incorrect segmentation. To leverage the uncertainty information, we propose a segmentation model that incorporates the uncertainty into its learning process. Firstly, we generate the uncertainty estimate (sample variance) using Monte-Carlo dropout during training. Then we incorporate it into the loss function t…
The impact of grain size on the efficiency of embedded SIMD image processing architectures
2004
Pixel-per-processing element (PPE) ratio-the amount of image data directly mapped to each processing element-has a significant impact on the area and energy efficiency of embedded SIMD architectures for image processing applications. This paper quantitatively evaluates the impact of PPE ratio on system performance and efficiency for focal-plane SIMD image processing architectures by comparing throughput, area efficiency, and energy efficiency for a range of common application kernels using architectural and workload simulation. While the impact of grain size is affected by the mix of executed instructions within an application program, the most efficient PPE ratio often does not occur at PE…
Space variant vision and pipelined architecture for time to impact computation
2002
Image analysis is one of the most interesting ways for a mobile vehicle to understand its environment. One of the tasks of an autonomous vehicle is to get accurate information of what it has in front, to avoid collision or find a way to a target. This task requires real-time restrictions depending on the vehicle speed and external object movement. The use of normal cameras, with homogeneous (squared) pixel distribution, for real-time image processing, usually requires high performance computing and high image rates. A different approach makes use of a CMOS space-variant camera that yields a high frame rate with low data bandwidth. The camera also performs the log-polar transform, simplifyin…
An FPGA-based design for real-time Super Resolution Reconstruction
2018
Since several decades, the camera spatial resolution is gradually increasing with the CMOS technology evolution. The image sensors provide more and more pixels, generating new constraints for the suitable optics. As an alternative, promising solutions propose Super Resolution (SR) image reconstruction to extend the image size without modifying the sensor architecture. Convincing state-of art studies demonstrate that these methods could even be implemented in real-time. Nevertheless, artifacts can be observed in highly textured areas of the image. In this paper, we propose a Local Adaptive Spatial Super Resolution (LASSR) method to fix this limitation. A real-time texture analysis is include…
Speeding-Up Differential Motion Detection Algorithms Using a Change-Driven Data Flow Processing Strategy
2007
A constraint of real-time implementation of differential motion detection algorithms is the large amount of data to be processed. Full image processing is usually the classical approach for these algorithms: spatial and temporal derivatives are calculated for all pixels in the image despite the fact that the majority of image pixels may not have changed from one frame to the next. By contrast, the data flow model works in a totally different way as instructions are only fired when the data needed for these instructions are available. Here we present a method to speed-up low level motion detection algorithms. This method is based on pixel change instead of full image processing and good spee…
A class-separability-based method for multi/hyperspectral image color visualization
2010
In this paper, a new color visualization technique for multi- and hyperspectral images is proposed. This method is based on a maximization of the perceptual distance between the scene endmembers as well as natural constancy of the resulting images. The stretched CMF principle is used to transform reflectance into values in the CIE L*a*b* colorspace combined with an a priori known segmentation map for separability enhancement between classes. Boundaries are set in the a*b* subspace to balance the natural palette of colors in order to ease interpretation by a human expert. Convincing results on two different images are shown.
Deep Learning-Based Sign Language Digits Recognition From Thermal Images With Edge Computing System
2021
The sign language digits based on hand gestures have been utilized in various applications such as human-computer interaction, robotics, health and medical systems, health assistive technologies, automotive user interfaces, crisis management and disaster relief, entertainment, and contactless communication in smart devices. The color and depth cameras are commonly deployed for hand gesture recognition, but the robust classification of hand gestures under varying illumination is still a challenging task. This work presents the design and deployment of a complete end-to-end edge computing system that can accurately provide the classification of hand gestures captured from thermal images. A th…