The PASCAL Visual Object Classes Challenge

While 3D object representations are being revived in the context of multi-view object class detection and scene understanding, they have not yet attained wide-spread use in fine-grained categorization. State-of-the-art approaches achieve remarkable performance when training data is plentiful, but they are typically tied to flat, 2D representations that model objects as a collection of unconnected views, limiting their ability to generalize across viewpoints. In this paper, we therefore lift two state-of-the-art 2D object representations to 3D, on the level of both local feature appearance and location. In extensive experiments on existing and newly proposed datasets, we show our 3D object representations outperform their state-of-the-art 2D counterparts for fine-grained categorization and demonstrate their efficacy for estimating 3D geometry from images via ultra-wide baseline matching and 3D reconstruction.

/pdf/3d-object-representations-for-fine-grained-categorization-2fpzpw4ir6.pdf

3D Object Representations for Fine-Grained Categorization

We explore the effect of dimensionality on the nearest neighbor problem. We show that under a broad set of conditions (much broader than independent and identically distributed dimensions), as dimensionality increases, the distance to the nearest data point approaches the distance to the farthest data point. To provide a practical perspective, we present empirical results on both real and synthetic data sets that demonstrate that this effect can occur for as few as 10-15 dimensions. These results should not be interpreted to mean that high-dimensional indexing is never meaningful; we illustrate this point by identifying some high-dimensional workloads for which this effect does not occur. However, our results do emphasize that the methodology used almost universally in the database literature to evaluate high-dimensional indexing techniques is flawed, and should be modified. In particular, most such techniques proposed in the literature are not evaluated versus simple linear scan, and are evaluated over workloads for which nearest neighbor is not meaningful. Often, even the reported experiments, when analyzed carefully, show that linear scan would outperform the techniques being proposed on the workloads studied in high (10-15) dimensionality!.

When is nearest neighbor meaningful

For the last two decades, intelligent transportation systems (ITS) have emerged as an efficient way of improving the performance of transportation systems, enhancing travel security, and providing more choices to travelers. A significant change in ITS in recent years is that much more data are collected from a variety of sources and can be processed into various forms for different stakeholders. The availability of a large amount of data can potentially lead to a revolution in ITS development, changing an ITS from a conventional technology-driven system into a more powerful multifunctional data-driven intelligent transportation system (D2ITS) : a system that is vision, multisource, and learning algorithm driven to optimize its performance. Furthermore, D2ITS is trending to become a privacy-aware people-centric more intelligent system. In this paper, we provide a survey on the development of D2ITS, discussing the functionality of its key components and some deployment issues associated with D2ITS Future research directions for the development of D2ITS is also presented.

Data-Driven Intelligent Transportation Systems: A Survey

We propose a network for Congested Scene Recognition called CSRNet to provide a data-driven and deep learning method that can understand highly congested scenes and perform accurate count estimation as well as present high-quality density maps. The proposed CSRNet is composed of two major components: a convolutional neural network (CNN) as the front-end for 2D feature extraction and a dilated CNN for the back-end, which uses dilated kernels to deliver larger reception fields and to replace pooling operations. CSRNet is an easy-trained model because of its pure convolutional structure. We demonstrate CSRNet on four datasets (ShanghaiTech dataset, the UCF_CC_50 dataset, the WorldEXPO'10 dataset, and the UCSD dataset) and we deliver the state-of-the-art performance. In the ShanghaiTech Part_B dataset, CSRNet achieves 47.3% lower Mean Absolute Error (MAE) than the previous state-of-the-art method. We extend the targeted applications for counting other objects, such as the vehicle in TRANCOS dataset. Results show that CSRNet significantly improves the output quality with 15.4% lower MAE than the previous state-of-the-art approach.

/pdf/csrnet-dilated-convolutional-neural-networks-for-5cdww8hhko.pdf

CSRNet: Dilated Convolutional Neural Networks for Understanding the Highly Congested Scenes

This paper presents an automatic road-sign detection and recognition system based on support vector machines (SVMs). In automatic traffic-sign maintenance and in a visual driver-assistance system, road-sign detection and recognition are two of the most important functions. Our system is able to detect and recognize circular, rectangular, triangular, and octagonal signs and, hence, covers all existing Spanish traffic-sign shapes. Road signs provide drivers important information and help them to drive more safely and more easily by guiding and warning them and thus regulating their actions. The proposed recognition system is based on the generalization properties of SVMs. The system consists of three stages: 1) segmentation according to the color of the pixel; 2) traffic-sign detection by shape classification using linear SVMs; and 3) content recognition based on Gaussian-kernel SVMs. Because of the used segmentation stage by red, blue, yellow, white, or combinations of these colors, all traffic signs can be detected, and some of them can be detected by several colors. Results show a high success rate and a very low amount of false positives in the final recognition stage. From these results, we can conclude that the proposed algorithm is invariant to translation, rotation, scale, and, in many situations, even to partial occlusions

/pdf/road-sign-detection-and-recognition-based-on-support-vector-3indwhzhzq.pdf

Road-Sign Detection and Recognition Based on Support Vector Machines

The challenging problem that we explore in this paper is to precisely estimate the number of vehicles in an image of a traffic congestion situation We start introducing TRANCOS, a novel database for extremely overlapping vehicle counting It provides more than 1200 images where the number of vehicles and their locations have been annotated We establish a clear experimental setup which will let others evaluate their own vehicle counting approaches We also propose a novel evaluation metric, the Grid Average Mean absolute Error (GAME), which overcomes the limitations of previously proposed metrics for object counting Finally, we perform an experimental validation, using the proposed TRANCOS dataset, for two types of vehicle counting strategies: counting by detection; and counting by regression Our results show that counting by regression strategies are more precise localizing and estimating the number of vehicles The TRANCOS database and the source code for reproducing the results are available at http://agamenontscuahes/Personales/rlopez/data/trancos

Extremely Overlapping Vehicle Counting

This paper presents a quantitative comparison of several segmentation methods (including new ones) that have successfully been used in traffic sign recognition The methods presented can be classified into color-space thresholding, edge detection, and chromatic/achromatic decomposition Our support vector machine (SVM) segmentation method and speed enhancement using a lookup table (LUT) have also been tested The best algorithm will be the one that yields the best global results throughout the whole recognition process, which comprises three stages: 1) segmentation; 2) detection; and 3) recognition Thus, an evaluation method, which consists of applying the entire recognition system to a set of images with at least one traffic sign, is attempted while changing the segmentation method used This way, it is possible to observe modifications in performance due to the kind of segmentation used The results lead us to conclude that the best methods are those that are normalized with respect to illumination, such as RGB or Ohta Normalized, and there is no improvement in the use of Hue Saturation Intensity (HSI)-like spaces In addition, an LUT with a reduction in the less-significant bits, such as that proposed here, improves speed while maintaining quality SVMs used in color segmentation give good results, but some improvements are needed when applied to achromatic colors

Goal Evaluation of Segmentation Algorithms for Traffic Sign Recognition

We present a new method to design prototype filters for conventional cosine-modulated pseudo-quadrature mirror filter (QMF) banks. This method is based on windowing, and sets the 3-dB cutoff frequency of the filter obtained at /spl pi//2M. In this way, the filter bank performance can be significantly improved compared to other existing design methods.

http://yli-kaakinen.fi/publications/EUSIPCO00/Cruz-Roldan,%20F.%20(2002,%20SPL)%20An%20Efficient%20and%20Simple%20Method%20for%20Designing%20Prototype%20Filters%20for%20Cosine-Modulated%20Pseudo-QMF%20Banks.pdf

An efficient and simple method for designing prototype filters for cosine-modulated pseudo-QMF banks

This paper deals with the detection and classification of traffic signs in outdoor environments. The information provided by traffic signs on roads is very important for the safety of drivers. However, in these situations the illumination conditions can not be predicted, the position and the orientation of signs in the scene are not known and other objects can block the vision of them. For these reasons we have developed an extensive test set which includes all kind of signs. In an artificial vision system, the key to recognize traffic signs is how to detect them and identify their geometric shapes. So, in this work we propose a method that uses a technique based on support vector machines (SVMs) for the classification. The patterns generated by the vectors represent the distances to borders (DtB) of the objects candidate to be traffic signs. Experimental results show the effectiveness of the proposed method.

Saturnino Maldonado-Bascón

Papers

Road-Sign Detection and Recognition Based on Support Vector Machines

Extremely Overlapping Vehicle Counting

Goal Evaluation of Segmentation Algorithms for Traffic Sign Recognition

An efficient and simple method for designing prototype filters for cosine-modulated pseudo-QMF banks

Traffic sign shape classification evaluation I: SVM using distance to borders