86,698 research outputs found
A survey on modern trainable activation functions
No full textIn neural networks literature, there is a strong interest in identifying and defining activation functions which can improve neural network performance. In recent years there has been a renovated interest in the scientific community in investigating activation functions which can be trained during the learning process, usually referred to as trainable, learnable or adaptable activation functions. They appear to lead to better network performance. Diverse and heterogeneous models of trainable activation function have been proposed in the literature. In this paper, we present a survey of these models. Starting from a discussion on the use of the term “activation function” in literature, we propose a taxonomy of trainable activation functions, highlight common and distinctive proprieties of recent and past models, and discuss main advantages and limitations of this type of approach. We show that many of the proposed approaches are equivalent to adding neuron layers which use fixed (non-trainable) activation functions and some simple local rule that constrains the corresponding weight layers
A distributed genetic algorithm for restoration of vertical line scratches
No full textThis paper reports a distributed algorithm for the restoration of still frames corrupted by
vertical line scratches. The restoration is here approached as an optimisation problem,
and is solved using an ad-hoc Genetic Algorithm. The distributed algorithm is designed following
a pipeline logical structure. The front end is a network of standard workstations
with heterogeneous operating systems. The quality of image is appreciable and the computational
time is quite low with respect the sequential version
A genetic algorithm for scratch removal from static images
No full textThis paper investigates the removal of line scratches from old moving pictures and gives a twofold contribution. First, it presents a simple technique for detecting the scratches, based on an analysis of the statistics of the grey levels. Second, the scratch removal is approached as an optimisation problem, which is solved by using a genetic algorithm. The method can be classified as a static approach, as it works independently on each single frame of the sequence. It does not require any a-priori knowledge of the absolute position of the scratch, nor an external starting population of chromosomes for the genetic algorithm. The central column of the line scratch once detected is changed with a conventional linear interpolation; this transformation is the starting point of the optimisation process.
Silhouette encoding and synthesis using elliptic Fourier descriptors and applications to videoconferencing, Journal of Visual Language and Computing
No full textThis paper investigates the use of elliptic Fourier descriptors as a shape descriptor for
encoding the silhouette of a person. Shape descriptors are here used for predicting the shape of
silhouettes in missing frames within a sequence. This prediction scheme is applied to the case
of generating in-between images in a low frame rate videoconferencing system, where the
reconstructed silhouette is used as a binary mask for reducing the computational time for the
frame reconstruction
Improving face recognition in low quality video sequences: Single frame vs multi-frame super-resolution
No full textRe-Identification aims to detect the presence of a subject spotted in one video in other videos. Traditional methods use information extracted from single frames like color, clothes, etc. A sequence in time domain of consecutive subject images could contain a greater amount of information compared with a single image of the same subject. Typically, these sequences are taken from surveillance cameras at very poor resolution. Even with modern cameras the resolution can be a problem when dealing with a subject who is far from the camera. A possible way of handling low resolution images is by using a multi-frame super-resolution algorithm. Multi-frame super-resolution image reconstruction aims at obtaining a high-resolution image by fusing a set of low-resolution images. Low-resolution images are usually subject to some degradation which causes substantial information loss. Therefore, contiguous images in a sequence could be viewed as a degraded version (SR image) of an image at higher resolution (HR image). Using a multi-frame SR algorithm could achieve a restoration of the HR image. This work aims to investigate the possibility of using a multi-frame super-resolution algorithm to enhance the performance of a classic re-identification system by exploiting information provided by video sequences made available by a video surveillance system. In the case that the SR technique employed results in an effective performance enhancement, we intend to show empirically how many match frames are required to have an effective improvement
XAI approach for addressing the dataset shift problem: BCI as a case study
No full textIn the Machine Learning (ML) literature, a well-known problem is the Dataset Shift problem where, differently from the ML standard hypothesis, the data in the training and test sets can follow different probability distributions leading ML systems toward poor generalisation performances. Therefore, such systems can be unreliable and risky, particularly when used in safety-critical domains. This problem is intensely felt in the Brain-Computer Interface (BCI) context, where bio-signals as Electroencephalographic (EEG) are used. In fact, EEG signals are highly non-stationary signals both over time and between different subjects. Despite several efforts in developing BCI systems to deal with different acquisition times or subjects, performance in many BCI applications remains low. Exploiting the knowledge from eXplainable Artificial Intelligence (XAI) methods can help develop EEG-based AI approaches, overcoming the performance returned by the current ones. The proposed framework will give greater robustness and reliability to BCI systems with respect to the current state of the art, alleviating the dataset shift problem and allowing a BCI system to be used by different subjects at different times without the need for further calibration/training stages
A simple and efficient architecture for trainable activation functions
No full textAutomatically learning the best activation function for the task is an active topic in neural network research. At the moment, despite promising results, it is still challenging to determine a method for learning an activation function that is, at the same time, theoretically simple and easy to implement. Moreover, most of the methods proposed so far introduce new parameters or adopt different learning techniques. In this work, we propose a simple method to obtain a trained activation function which adds to the neural network local sub-networks with a small number of neurons. Experiments show that this approach could lead to better results than using a pre-defined activation function, without introducing the need to learn a large number of additional parameters
On projective rectification
No full textWe present a novel algorithm performing projective rectification which does not require explicit computation of the epipolar geometry, and specifically of the fundamental matrix. Instead of finding the epipoles and computing two homographies mapping the epipoles to infinity, as done in recent work on projective rectification, we exploit the fact that the fundamental matrix of a pair of rectified images has a particular, known form. This allows us to set up a minimization that yields the rectifying homographies directly from image correspondences. Experimental results show that our method works quite robustly even in the presence of noise, and can cope with inaccurate point correspondences.</p
- …
