In the field of Art History, images of artworks and their contexts are core to understanding the underlying semantic information. However, the highly complex and sophisticated representation of these artworks makes it difficult, even for the experts, to analyze the scene. From the computer vision perspective, the task of analyzing such artworks can be divided into sub-problems by taking a bottom-up approach. In this paper, we focus on the problem of recognizing the characters in Art History. From the iconography of $Annunciation$ $of$ $the$ $Lord$ (Figure 1), we consider the representation of the main protagonists, $Mary$ and $Gabriel$, across different artworks and styles. We investigate and present the findings of training a character classifier on features extracted from their face images. The limitations of this method, and the inherent ambiguity in the representation of $Gabriel$, motivated us to consider their bodies (a bigger context) to analyze in order to recognize the characters. Convolutional Neural Networks (CNN) trained on the bodies of $Mary$ and $Gabriel$ are able to learn person related features and ultimately improve the performance of character recognition. We introduce a new technique that generates more data with similar styles, effectively creating data in the similar domain. We present experiments and analysis on three different models and show that the model trained on domain related data gives the best performance for recognizing character. Additionally, we analyze the localized image regions for the network predictions. Code is open-sourced and available at https://github.com/prathmeshrmadhu/recognize_characters_art_history and the link to the published peer-reviewed article is https://dl.acm.org/citation.cfm?id=3357242.
Most people think that their handwriting is unique and cannot be imitated by machines, especially not using completely new content. Current cursive handwriting synthesis is visually limited or needs user interaction. We show that subdividing the process into smaller subtasks makes it possible to imitate someone's handwriting with a high chance to be visually indistinguishable for humans. Therefore, a given handwritten sample will be used as the target style. This sample is transferred to an online sequence. Then, a method for online handwriting synthesis is used to produce a new realistic-looking text primed with the online input sequence. This new text is then rendered and style-adapted to the input pen. We show the effectiveness of the pipeline by generating in- and out-of-vocabulary handwritten samples that are validated in a comprehensive user study. Additionally, we show that also a typical writer identification system can partially be fooled by the created fake handwritings.
Photovoltaic is one of the most important renewable energy sources for dealing with world-wide steadily increasing energy consumption. This raises the demand for fast and scalable automatic quality management during production and operation. However, the detection and segmentation of cracks on electroluminescence (EL) images of mono- or polycrystalline solar modules is a challenging task. In this work, we propose a weakly supervised learning strategy that only uses image-level annotations to obtain a method that is capable of segmenting cracks on EL images of solar cells. We use a modified ResNet-50 to derive a segmentation from network activation maps. We use defect classification as a surrogate task to train the network. To this end, we apply normalized Lp normalization to aggregate the activation maps into single scores for classification. In addition, we provide a study how different parameterizations of the normalized Lp layer affect the segmentation performance. This approach shows promising results for the given task. However, we think that the method has the potential to solve other weakly supervised segmentation problems as well.
This competition investigates the performance of large-scale retrieval of historical document images based on writing style. Based on large image data sets provided by cultural heritage institutions and digital libraries, providing a total of 20 000 document images representing about 10 000 writers, divided in three types: writers of (i) manuscript books, (ii) letters, (iii) charters and legal documents. We focus on the task of automatic image retrieval to simulate common scenarios of humanities research, such as writer retrieval. The most teams submitted traditional methods not using deep learning techniques. The competition results show that a combination of methods is outperforming single methods. Furthermore, letters are much more difficult to retrieve than manuscripts.
Global pooling layers are an essential part of Convolutional Neural Networks (CNN). They are used to aggregate activations of spatial locations to produce a fixed-size vector in several state-of-the-art CNNs. Global average pooling or global max pooling are commonly used for converting convolutional features of variable size images to a fix-sized embedding. However, both pooling layer types are computed spatially independent: each individual activation map is pooled and thus activations of different locations are pooled together. In contrast, we propose Deep Generalized Max Pooling that balances the contribution of all activations of a spatially coherent region by re-weighting all descriptors so that the impact of frequent and rare ones is equalized. We show that this layer is superior to both average and max pooling on the classification of Latin medieval manuscripts (CLAMM'16, CLAMM'17), as well as writer identification (Historical-WI'17).
Fast, non-destructive and on-site quality control tools, mainly high sensitive imaging techniques, are important to assess the reliability of photovoltaic plants. To minimize the risk of further damages and electrical yield losses, electroluminescence (EL) imaging is used to detect local defects in an early stage, which might cause future electric losses. For an automated defect recognition on EL measurements, a robust detection and rectification of modules, as well as an optional segmentation into cells is required. This paper introduces a method to detect solar modules and crossing points between solar cells in EL images. We only require 1-D image statistics for the detection, resulting in an approach that is computationally efficient. In addition, the method is able to detect the modules under perspective distortion and in scenarios, where multiple modules are visible in the image. We compare our method to the state of the art and show that it is superior in presence of perspective distortion while the performance on images, where the module is roughly coplanar to the detector, is similar to the reference method. Finally, we show that we greatly improve in terms of computational time in comparison to the reference method.
Accurate noise modelling is important for training of deep learning reconstruction algorithms. While noise models are well known for traditional imaging techniques, the noise distribution of a novel sensor may be difficult to determine a priori. Therefore, we propose learning arbitrary noise distributions. To do so, this paper proposes a fully connected neural network model to map samples from a uniform distribution to samples of any explicitly known probability density function. During the training, the Jensen-Shannon divergence between the distribution of the model's output and the target distribution is minimized. We experimentally demonstrate that our model converges towards the desired state. It provides an alternative to existing sampling methods such as inversion sampling, rejection sampling, Gaussian mixture models and Markov-Chain-Monte-Carlo. Our model has high sampling efficiency and is easily applied to any probability distribution, without the need of further analytical or numerical calculations.
Electroluminescence (EL) imaging is a useful modality for the inspection of photovoltaic (PV) modules. EL images provide high spatial resolution, which makes it possible to detect even finest defects on the surface of PV modules. However, the analysis of EL images is typically a manual process that is expensive, time-consuming, and requires expert knowledge of many different types of defects. In this work, we investigate two approaches for automatic detection of such defects in a single image of a PV cell. The approaches differ in their hardware requirements, which are dictated by their respective application scenarios. The more hardware-efficient approach is based on hand-crafted features that are classified in a Support Vector Machine (SVM). To obtain a strong performance, we investigate and compare various processing variants. The more hardware-demanding approach uses an end-to-end deep Convolutional Neural Network (CNN) that runs on a Graphics Processing Unit (GPU). Both approaches are trained on 1,968 cells extracted from high resolution EL intensity images of mono- and polycrystalline PV modules. The CNN is more accurate, and reaches an average accuracy of 88.42%. The SVM achieves a slightly lower average accuracy of 82.44%, but can run on arbitrary hardware. Both automated approaches make continuous, highly accurate monitoring of PV cells feasible.
Deep learning approaches have shown promising performance for compressed sensing-based Magnetic Resonance Imaging. While deep neural networks trained with mean squared error (MSE) loss functions can achieve high peak signal to noise ratio, the reconstructed images are often blurry and lack sharp details, especially for higher undersampling rates. Recently, adversarial and perceptual loss functions have been shown to achieve more visually appealing results. However, it remains an open question how to (1) optimally combine these loss functions with the MSE loss function and (2) evaluate such a perceptual enhancement. In this work, we propose a hybrid method, in which a visual refinement component is learnt on top of an MSE loss-based reconstruction network. In addition, we introduce a semantic interpretability score, measuring the visibility of the region of interest in both ground truth and reconstructed images, which allows us to objectively quantify the usefulness of the image quality for image post-processing and analysis. Applied on a large cardiac MRI dataset simulated with 8-fold undersampling, we demonstrate significant improvements ($p<0.01$) over the state-of-the-art in both a human observer study and the semantic interpretability score.
Embedding data into vector spaces is a very popular strategy of pattern recognition methods. When distances between embeddings are quantized, performance metrics become ambiguous. In this paper, we present an analysis of the ambiguity quantized distances introduce and provide bounds on the effect. We demonstrate that it can have a measurable effect in empirical data in state-of-the-art systems. We also approach the phenomenon from a computer security perspective and demonstrate how someone being evaluated by a third party can exploit this ambiguity and greatly outperform a random predictor without even access to the input data. We also suggest a simple solution making the performance metrics, which rely on ranking, totally deterministic and impervious to such exploits.