Multilayer neural networks trained with the back-propagation algorithm constitute the best example of a successful gradient based learning technique. Given an appropriate network architecture, gradient-based learning algorithms can be used to synthesize a complex decision surface that can classify high-dimensional patterns, such as handwritten characters, with minimal preprocessing. This paper reviews various methods applied to handwritten character recognition and compares them on a standard handwritten digit recognition task. Convolutional neural networks, which are specifically designed to deal with the variability of 2D shapes, are shown to outperform all other techniques. Real-life document recognition systems are composed of multiple modules including field extraction, segmentation recognition, and language modeling. A new learning paradigm, called graph transformer networks (GTN), allows such multimodule systems to be trained globally using gradient-based methods so as to minimize an overall performance measure. Two systems for online handwriting recognition are described. Experiments demonstrate the advantage of global training, and the flexibility of graph transformer networks. A graph transformer network for reading a bank cheque is also described. It uses convolutional neural network character recognizers combined with global training techniques to provide record accuracy on business and personal cheques. It is deployed commercially and reads several million cheques per day.

Gradient-based learning applied to document recognition

We propose a deep convolutional neural network architecture codenamed Inception that achieves the new state of the art for classification and detection in the ImageNet Large-Scale Visual Recognition Challenge 2014 (ILSVRC14). The main hallmark of this architecture is the improved utilization of the computing resources inside the network. By a carefully crafted design, we increased the depth and width of the network while keeping the computational budget constant. To optimize quality, the architectural decisions were based on the Hebbian principle and the intuition of multi-scale processing. One particular incarnation used in our submission for ILSVRC14 is called GoogLeNet, a 22 layers deep network, the quality of which is assessed in the context of classification and detection.

/pdf/going-deeper-with-convolutions-1yobw2o2ds.pdf

Going deeper with convolutions

There is, I think, something ethereal about i —the square root of minus one. I remember first hearing about it at school. It seemed an odd beast at that time—an intruder hovering on the edge of reality.

Usually familiarity dulls this sense of the bizarre, but in the case of i it was the reverse: over the years the sense of its surreal nature intensified. It seemed that it was impossible to write mathematics that described the real world in …

I and i

We study the question of feature sets for robust visual object recognition; adopting linear SVM based human detection as a test case. After reviewing existing edge and gradient based descriptors, we show experimentally that grids of histograms of oriented gradient (HOG) descriptors significantly outperform existing feature sets for human detection. We study the influence of each stage of the computation on performance, concluding that fine-scale gradients, fine orientation binning, relatively coarse spatial binning, and high-quality local contrast normalization in overlapping descriptor blocks are all important for good results. The new approach gives near-perfect separation on the original MIT pedestrian database, so we introduce a more challenging dataset containing over 1800 annotated human images with a large range of pose variations and backgrounds.

/pdf/histograms-of-oriented-gradients-for-human-detection-3nteinziuq.pdf

Histograms of oriented gradients for human detection

Object detection performance, as measured on the canonical PASCAL VOC dataset, has plateaued in the last few years. The best-performing methods are complex ensemble systems that typically combine multiple low-level image features with high-level context. In this paper, we propose a simple and scalable detection algorithm that improves mean average precision (mAP) by more than 30% relative to the previous best result on VOC 2012 -- achieving a mAP of 53.3%. Our approach combines two key insights: (1) one can apply high-capacity convolutional neural networks (CNNs) to bottom-up region proposals in order to localize and segment objects and (2) when labeled training data is scarce, supervised pre-training for an auxiliary task, followed by domain-specific fine-tuning, yields a significant performance boost. Since we combine region proposals with CNNs, we call our method R-CNN: Regions with CNN features. We also present experiments that provide insight into what the network learns, revealing a rich hierarchy of image features. Source code for the complete system is available at http://www.cs.berkeley.edu/~rbg/rcnn.

/pdf/rich-feature-hierarchies-for-accurate-object-detection-and-22fvjlggyg.pdf

Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation

Vision-Based Autonomous Helicopter Research at Carnegie Mellon Robotics Institute 1991-1997

An iterative stereo matching algorithm is presented which selects a window adaptively for each pixel. The selected window is optimal in the sense that it produces the disparity estimate having the least uncertainty after evaluating both the intensity and the disparity variations within a window. The algorithm employs a statistical model that represents uncertainty of disparity of points over the window; the uncertainty is assumed to increase with the distance of the point from the center point. The algorithm is completely local and does not include any global optimization. Also, the algorithm does not use any post-processing smoothing, but smooth surfaces are recovered as smooth while sharp disparity edges are retained. Experimental results have demonstrated a clear advantage of this algorithm over algorithms with a fixed-size window, for both synthetic and real images. >

/pdf/a-stereo-matching-algorithm-with-an-adaptive-window-theory-1yi0mrudl3.pdf

A stereo matching algorithm with an adaptive window: theory and experiment

Representing images with layers has many important applications, such as video compression, motion analysis, and 3D scene analysis. This paper presents an approach to reliably extracting layers from images by taking advantages of the fact that homographies induced by planar patches in the scene form a low dimensional linear subspace. Layers in the input images will be mapped in the subspace, where it is proven that they form well-defined clusters and can be reliably identified by a simple mean-shift based clustering algorithm. Global optimality is achieved since all valid regions are simultaneously taken into account, and noise can be effectively reduced by enforcing the subspace constraint. Good layer descriptions are shown to be extracted in the experimental results.

/pdf/a-subspace-approach-to-layer-extraction-3je84o67br.pdf

A subspace approach to layer extraction

Automated cell tracking in populations is important for research and discovery in biology and medicine. In this paper, we propose a cell tracking method based on global spatio-temporal data association which considers hypotheses of initialization, termination, translation, division and false positive in an integrated formulation. Firstly, reliable tracklets (i.e., short trajectories) are generated by linking detection responses based on frame-by-frame association. Next, these tracklets are globally associated over time to obtain final cell trajectories and lineage trees. During global association, tracklets form tree structures where a mother cell divides into two daughter cells. We formulate the global association for tree structures as a maximum-a-posteriori (MAP) problem and solve it by linear programming. This approach is quantitatively evaluated on sequences with thousands of cells captured over several days.

/pdf/reliable-cell-tracking-by-global-data-association-damfxdb6p8.pdf

Reliable cell tracking by global data association

We present a fast pattern matching algorithm with a large set of templates. The algorithm is based on the typical template matching speeded up by the dual decomposition; the Fourier transform and the Karhunen-Loeve transform. The proposed algorithm is appropriate for the search of an object with unknown distortion within a short period. Patterns with different distortion differ slightly from each other and are highly correlated. The image vector subspace required for effective representation can be defined by a small number of eigenvectors derived by the Karhunen-Loeve transform. A vector subspace spanned by the eigenvectors is generated, and any image vector in the subspace is considered as a pattern to be recognized. The pattern matching of objects with unknown distortion is formulated as the process to extract the portion of the input image, find the pattern most similar to the extracted portion in the subspace, compute normalized correlation between them at each location in the input image, and find the location with the best score. Searching for objects with unknown distortion requires vast computation. The formulation above makes it possible to decompose highly correlated reference images into eigenvectors, as well as to decompose images in frequency domain, and to speed up the process significantly.

/pdf/use-of-fourier-and-karhunen-loeve-decomposition-for-fast-4l9tgcz5q3.pdf

Takeo Kanade

Papers

Vision-Based Autonomous Helicopter Research at Carnegie Mellon Robotics Institute 1991-1997

A stereo matching algorithm with an adaptive window: theory and experiment

A subspace approach to layer extraction

Reliable cell tracking by global data association

Use of Fourier and Karhunen-Loeve decomposition for fast pattern matching with a large set of templates