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

The Carnegie—Mellon University Navigational Laboratory (the CMU Navlab) project integrates sensing, image understanding, planning, control, and software systems architectures onto a self-contained mobile robot. The Navlab drives autonomously along a variety of roads (dirt, gravel unmarked bicycle paths, city streets, rural roads) and cross-country. This chapter describes the Navlab and its contributions in color vision, neural nets, 3-D perception, planning, and robot architectures.

The new generation system for the CMU Navlab

We propose a data-driven approach to estimate the likelihood that an image segment corresponds to a scene object (its “objectness”) by comparing it to a large collection of example object regions. We demonstrate that when the application domain is known, for example, in our case activity of daily living (ADL), we can capture the regularity of the domain specific objects using millions of exemplar object regions. Our approach to estimating the objectness of an image region proceeds in two steps: $1)$ finding the exemplar regions that are the most similar to the input image segment; $2)$ calculating the objectness of the image segment by combining segment properties, mutual consistency across the nearest exemplar regions, and the prior probability of each exemplar region. In previous work, parametric objectness models were built from a small number of manually annotated objects regions, instead, our data-driven approach uses $5$ million object regions along with their metadata information. Results on multiple data sets demonstrates our data-driven approach compared to the existing model based techniques. We also show the application of our approach in improving the performance of object discovery algorithms.

Data-Driven Objectness

Knowing how well an activity is performed is important for home rehabilitation. We would like to not only know if a motion is being performed correctly, but also in what way the motion is incorrect so that we may provide feedback to the user. This paper describes methods for assessing human motion quality using body-worn tri-axial accelerometers and gyroscopes. We use multi-label classifiers to detect subtle errors in exercise performances of eight individuals with knee osteoarthritis, a degenerative disease of the cartilage. We present results obtained using various machine learning methods with decision tree base classifiers. The classifier can detect classes in multi-label data with 75% sensitivity, 90% specificity and 80% accuracy. The methods presented here form the basis for an at-home rehabilitation device that will recognize errors in patient exercise performance, provide appropriate feedback on the performance, and motivate the patient to continue the prescribed regimen.

Multi-label classification for the analysis of human motion quality

In this paper, we propose an approach for construction of a projective 3D voxel space based on epipolar geometry obtained with weak calibration. This concept of voxel space defines the epipolar lines following to the coordinate axis. In the field of computer vision, it is common to reconstruct 3D geometry data based on camera calibration data and disparities of matching points in each image. In the case that the goal of the system is generating the images from another point of view, complete 3D geometry reconstruction is not presently required. However, detecting the consistent matching points in several pairs of images without complete 3D geometry information is difficult. This paper describes the concept of projective 3D voxel, which makes it possible to handle 3D geometric data without complete 3D geometry information.

3D voxel construction based on epipolar geometry

Appearance models (AM) are commonly used to model appearance and shape variation of objects in images. In particular, they have proven useful to detection, tracking, and synthesis of people's faces from video. While AM have numerous advantages relative to alternative approaches, they have at least two important drawbacks. First, they are especially prone to local minima in fitting; this problem becomes increasingly problematic as the number of parameters to estimate grows. Second, often few if any of the local minima correspond to the correct location of the model error. To address these problems, we propose filtered component analysis (FCA), an extension of traditional principal component analysis (PCA). FCA learns an optimal set of filters with which to build a multi-band representation of the object. FCA representations were found to be more robust than either grayscale or Gabor filters to problems of local minima. The effectiveness and robustness of the proposed algorithm is demonstrated in both synthetic and real data.

/pdf/filtered-component-analysis-to-increase-robustness-to-local-1rhe4tl9fl.pdf

Takeo Kanade

Papers

The new generation system for the CMU Navlab

Data-Driven Objectness

Multi-label classification for the analysis of human motion quality

3D voxel construction based on epipolar geometry

Filtered Component Analysis to Increase Robustness to Local Minima in Appearance Models