International Symposium on Visual Computing 

About: International Symposium on Visual Computing is an academic conference. The conference publishes majorly in the area(s): Segmentation & Image processing. Over the lifetime, 2471 publications have been published by the conference receiving 19180 citations.

Optimizing Intersection-Over-Union in Deep Neural Networks for Image Segmentation

Abstract: We consider the problem of learning deep neural networks (DNNs) for object category segmentation, where the goal is to label each pixel in an image as being part of a given object (foreground) or not (background). Deep neural networks are usually trained with simple loss functions (e.g., softmax loss). These loss functions are appropriate for standard classification problems where the performance is measured by the overall classification accuracy. For object category segmentation, the two classes (foreground and background) are very imbalanced. The intersection-over-union (IoU) is usually used to measure the performance of any object category segmentation method. In this paper, we propose an approach for directly optimizing this IoU measure in deep neural networks. Our experimental results on two object category segmentation datasets demonstrate that our approach outperforms DNNs trained with standard softmax loss.

BlenSor: blender sensor simulation toolbox

Abstract: This paper introduces a novel software package for the simulation of various types of range scanners The goal is to provide researchers in the fields of obstacle detection, range data segmentation, obstacle tracking or surface reconstruction with a versatile and powerful software package that is easy to use and allows to focus on algorithmic improvements rather than on building the software framework around it The simulation environment and the actual simulations can be efficiently distributed with a single compact file Our proposed approach facilitates easy regeneration of published results, hereby highlighting the value of reproducible research

Continuous Pain Intensity Estimation from Facial Expressions

Abstract: Automatic pain recognition is an evolving research area with promising applications in health care. In this paper, we propose the first fully automatic approach to continuous pain intensity estimation from facial images. We first learn a set of independent regression functions for continuous pain intensity estimation using different shape (facial landmarks) and appearance (DCT and LBP) features, and then perform their late fusion. We show on the recently published UNBC-MacMaster Shoulder Pain Expression Archive Database that late fusion of the afore-mentioned features leads to better pain intensity estimation compared to feature-specific pain intensity estimation.

Convolutional Neural Networks for Steel Surface Defect Detection from Photometric Stereo Images

Abstract: Convolutional neural networks (CNNs) achieved impressive recognition rates in image classification tasks recently. In order to exploit those capabilities, we trained CNNs on a database of photometric stereo images of metal surface defects, i.e. rail defects. Those defects are cavities in the rail surface and are indication for further surface degradation right up to rail break. Due to security issues, defects have to be recognized early in order to take countermeasures in time. By means of differently colored light-sources illuminating the rail surfaces from different and constant directions, those cavities are made visible in a photometric dark-field setup. So far, a model-based approach has been used for image classification, which expressed the expected reflection properties of surface defects in contrast to non-defects. In this work, we experimented with classical CNNs trained in pure supervised manner and also explored the impact of regularization methods such as unsupervised layer-wise pre-training and training data-set augmentation. The classical CNN already distinctly outperforms the model-based approach. Moreover, regularization methods yet yield further improvements.

Diverging Color Maps for Scientific Visualization

Abstract: One of the most fundamental features of scientific visualization is the process of mapping scalar values to colors. This process allows us to view scalar fields by coloring surfaces and volumes. Unfortunately, the majority of scientific visualization tools still use a color map that is famous for its ineffectiveness: the rainbow color map. This color map, which naively sweeps through the most saturated colors, is well known for its ability to obscure data, introduce artifacts, and confuse users. Although many alternate color maps have been proposed, none have achieved widespread adoption by the visualization community for scientific visualization. This paper explores the use of diverging color maps (sometimes also called ratio, bipolar, or double-ended color maps) for use in scientific visualization, provides a diverging color map that generally performs well in scientific visualization applications, and presents an algorithm that allows users to easily generate their own customized color maps.